Binocular Facilitation and Inhibition in the Lateral Geniculate Nucleus of the Cat Citation Marton, Kenneth Lawrence (1980) Binocular Facilitation and Inhibition in the Lateral Geniculate Nucleus of the Cat. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/w5yj-2531. https://resolver.caltech.edu/CaltechTHESIS:10292025-013330758 Abstract The lateral geniculate nucleus (LGN) relays visual information from the retinas to the cortex, segregating input from each eye into separate laminae. The LGN receives an equally large input back from the visual cortex, whose cells are driven from both eyes. Therefore, binocular interactions in the LGN were studied by systematically varying visual stimuli known to fire cortical neurons. Binocular to monocular responses were compared by interleaving them using computer driven shutters in order to eliminate errors due to LGN cell response variability. Full statistical analysis was used to identify significant binocular facilitation and inhibition. Significantly more and stronger binocular feedback (BF) was seen with this approach than in previous studies. The vast majority of LGN cells showed both binocular facilitation and inhibition; as many as half showed BF amplitudes exceeding 50% of their typical monocular firing rate. Importantly, BF was found to be well tuned to velocity, relative retinal disparity, and sweep direction, parameters known to profoundly affect cortical firing. Multiple regions of BF were found for most cells, with the majority located near the monocular receptive fields in visual space. Regions of facilitation required zero retinal disparity twice as often as inhibition. Further, most BF reached maximum amplitudes at 6°/sec to 12°/sec. These results are a strong indication that the BF is cortical in origin. It is likely that this BF has a role in highlighting visual features on the plane of fixation. Because BF is very sensitive to parameters of motion, it is also conceivable that it is involved in interpreting the visual signals generated by eyes constantly in motion. This and other possible roles of BF are discussed. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Neurobiology) Degree Grantor: California Institute of Technology Division: Biology Major Option: Neurobiology Thesis Availability: Public (worldwide access) Research Advisor(s): Pettigrew, John D. Thesis Committee: Pettigrew, John D. (chair) Konishi, Masakazu Van Essen, David C. Fender, Derek H. Allman, John Morgan Defense Date: 5 June 1980 Record Number: CaltechTHESIS:10292025-013330758 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10292025-013330758 DOI: 10.7907/w5yj-2531 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17735 Collection: CaltechTHESIS Deposited By: Ben Maggio Deposited On: 29 Oct 2025 16:52 Last Modified: 29 Oct 2025 17:06 Thesis Files PDF - Final Version See Usage Policy. 29MB Repository Staff Only: item control page

BINOCULAR FACILITATION AND INHIBITION IN THE LATERAL GENICULATE NUCLEUS OF THE CAT Thesis by Kenneth Lawrence Marton In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1980 (Submitted June S, 1980) @ - 1980 Kenneth Lawrence Marton All Rights Reserved - ii Acknowledgements I am ~ndebted to my thesis advisor, Jack Pettigrew, for his and strong support patience throughout my graduate education. He is an out- standing scientist and a free and creative thinker. I would also like to thank my doctoral committee, Drs. Allman, Fender, Konishi, and Van Essen for support in completing my degree. I would like to express my eased appreciation to some friends my way considerably. Dr. Marie Ary provided frequent technical assistance, proofreading and intense personal support. I am to who Drs. David Presti, this manuscript. grateful v.s. Ramachandran, and Hermes Bravo who read These and other members of our research group at Caltech always provided an environment that was supportive, intellectually exciting, and even a nice place to be at 3:00 A.M.! There are a large number of people whose support throughout. Gary Blasdel and Herb was important Adams were responsible for the design and construction of some very high quality instruments used in the work. The entire administrative staff of the Biology Division under the management of Michael Miranda must be acknowledged truly serving the needs of myself and other researchers. to thank Joy Hansen and the Humanities Division for for I also want allowing me to use their computer printer for typing this manuscript. This work was supported by NIH-PHS grants EY01909 and EY03291 and this manuscript was prepared with the assistance of the Jean Weigle Memorial Fund. - iii Abstract The lateral geniculate nucleus (LGN) relays visual the retinas to the cortex, segregating information input from each eye into separate laminae. The LGN receives an equally large input the from back from visual cortex, whose cells are driven from both eyes. Therefore, binocular interactions in the varying visual LGN were studied by systematically stimuli known to fire cortical neurons. Binocular to monocular responses were compared by interleaving them using computer driven shutters in order to eliminate errors due to LGN cell response variability. Full statistical analysis was used to identify signifi- cant binocular facilitation and inhibition. Significantly more and stronger binocular feedback (BF) was seen with this approach than in previous studies. The vast majority of LGN cells showed both binocular facilitation and inhibition; as half showed BF velocity, retinal disparity, and sweep direction, parameters known to profoundly affect cortical firing. Multiple regions of BF were for most as amplitudes exceeding 50% of their typical monocular firing rate. Importantly, BF was found to be well tuned to relative many found cells, with the majority located near the monocular recep- tive fields in visual space. Regions of facilitation retinal often disparity twice as required zero as inhibition. Further, most BF reached maximum amplitudes at 6°/sec to 12°/sec. These results are a strong indication that the BF is cortical in origin. It is likely that this BF has features on the a role in highlighting visual plane of fixation. Because BF is very sensitive to - iv parameters of motion, it is also conceivable that it is interpreting the visual signals generated by involved in eyes constantly in motion. This and other possible roles of BF are discussed. - V - Table of Contents Acknowledgements • • . . . . . .. . ii Abstract iii ..... Table of Contents V vi Table of Abbreviations • 1. Introduction. 1 2. Methods 8 2.1 Electrophysiology 8 2.2 Visual Stimulation and Protocol 2.3 Data Analysis 11 19 . . . . . .. . . . . . . . . . . 3. Results 3.1 General 3.2 Location of Binocular Feedback in Visual Space 3.3 Velocity Tuning 3.4 Directional Selectivity 3.5 Other Observations Concerning Binocular Feedback 25 32 55 80 4.1 Relationship to Previous Work 98 4.2 Source of Binocular Feedback 101 4.3 Functional Significance 90 96 4. Discussion References •• 25 105 108 - vi Table of Abbreviations No - N degrees N°/sec - N degrees per second sis - spikes per second AC - Area Centralis BDH - Binocular Difference Histogram BF - Binocular Feedback CONTRA - Contralateral IPSI - Ipsilateral LGN - Lateral Geniculate Nucleus RF - Receptive Field - 1 1. INTRODUCTION By studying the anatomy and physiology of the dorsal lateral late nucleus (LGN), neurobiologists have gained an understanding of the role played by the LGN in transferring the retina to the cortex. visual from one information from In the cat and primate, this nucleus is strikingly divided into clear laminae, each lamina input genicu- receiving direct retina (Hubel and Wiesel, 1961; Kaas et al., 1972). Many properties of the relay cells in the LGN have been well studied and for the most part seem accounted for by a simple model consisting of each LGN cell being driven by one or a few retinal ganglion cells (Hubel and 1976). Recording simultaneously from pairs of retinal ganglion cells and LGN Wiesel, 1961; Bishop et al., 1962; Stevens and Gerstein, neurons, Cleland et al. (1971) showed that every spike from one LGN cell can be accounted for by the firing cells from one retina. of a few ganglion However, an LGN neuron produces fewer spikes than the retinal cells driving it and must by some mechanism select a subset of the spikes available to it. The Although simple the monocular model of LGN neurons is inadequate. evidence in favor of complete segregation of the path- ways from the two eyes in the LGN is strong, a number of studies have demonstrated binocular influences on LGN relay cells. Bishop and Davis (1953) first showed that a conditioning electric shock to one applied optic nerve depressed the post-synaptic field potential eli- cited by a test volley to the other nerve. Intracellular and single unit recording showed that electrical stimulation of the non-dominant - 2 eye's optic nerve would inhibit the firing of an LGN cell lows electric shocks that fol- to the dominant eye's nerve (Suzuki and Kato, 1966; Suzuki and Takahashi, 1970). The importance of binocular interactions in the functioning of the LGN was further supported by the discovery of inhibitory and excitatory fields from a cell's dominant eye (Singer, 1970; Sanderson et al., 1971). The fields were most often inhibitory and were discovered moving non- lines instead with of ·electric shocks. flashing spots and The presence of visually driven inhibition and occasionally facilitation of an LGN eell from the in awake, non-dominant eye has also been demonstrated chronically-implanted cats (Noda et al., 1972). that fields (RFs) have a center- some non-dominant LGN receptive There is evidence surround structure, with facilitation in the center and inhibition in the surround (Schmielau and Singer, 1977). All researchers agree that the fields in the non-dominant relative eye occupy approximately the same retinal position as the RF in the dominant eye and that the non-dominant responses are quite weak in comparison. To date, how- ever, the functional significance of these interactions is unclear. What could be responsible for binocular interactions in a structure that has such clear separation of the input from the two eyes? It is likely that the visual cortex is involved in these interactions in the LGN. The existence of a massive projection from the visual cortices to the LGN has been demonstrated repeatedly in both the and monkey. Lesions in areas 17, 18 and 19 of the cortex have been shown to result in synaptic degeneration in the LGN (Guillery, Kawamura et cat al., 1974). 1967; Retrograde transport and autoradiographic - 3 techniques have shown that the projection is heaviest from the stri- ate cortex and is topographically organized (Hollander, 1972; Gilbert and Kelly, 1975; Lund et al., 1975; Updyke, 1975). The projection very is large; more than half of the neurons in layer VI of the striate cortex send axons to the LGN (Gilbert and Kelly, 1975) and half at least of the synapses seen in the LGN are of the type that degenerate soon after the cortex is removed (Jones and Powell, 1971). Clearly, 1969; Guillery, this large feedback system from binocular cells in the visual cortex must profoundly affect LGN function. A number of workers have demonstrated that influence the cortex has on the activity of LGN relay cells. Widen and Marsan first showed in 1960 that an electric shock to the visual cortex can bit an or facilitate the response of an LGN inhi- unit to visual or electri- cal stimulation. Others have shown in both the cat and monkey that cooling the cortex decreases the spontaneous firing and the responses to visual stimuli in about one third of the LGN cells them in and increases about half that number (Hull, 1968; Kalil and Chase, 1970). Tsumoto et al. (1978) excited cortical neurons in small areas of layer VI in area 17 with the application of glutamate; both facilitation and inhibition of firing to visual stimuli could be found in LGN cells. Schmielau and Singer (1977) found that cortical cooling some- times altered the non-dominant LGN receptive fields; center facilitation was replaced by inhibition and surround inhibition was decre- mented. The implication that only some of the binocular inhibition in the LGN is mediated through the cortex is supported by the earlier studies of Singer (1970) and Sanderson et al. (1971) where it was - 4 shown that the non-dominant inhibitory cortical destruction or cooling. fields remained following The functional significance of this feedback remains to be elucidated. Because of the probable involvement study designed of the visual cortex, a to explore the nature and functional significance of binocular interactions in the LGN would best take properties of cortical neurons. into account the It is well established that cortical cells are binocular and respond best to moving oriented stimuli, with each cell expressing strong preferences for stimulus orientation, sweep velocity and direction, retinal disparity, as well as other parameters (Hubel and Wiesel, 1962; Barlow et al., 1967; Pettigrew et al., 1968a). It has already been shown that some LGN cells show weak preferences for orientation and it is possible that this is due to cortical influences (Daniels et al., 1977). optimal to consider the the LGN. To date, identification of these cells by antidromic firing from LGN reveals only that they can be either simple or complex the specific It would be properties of cortical cells that project to and are most often binocular (Gilbert, 1977; Harvey, 1978). Another issue that such a study should handle is the problem LGN unit response variability. of The amplitude of the response of an LGN neuron to a visual stimulus often varies as much as more with factors that control (Malcolm et al., 1970; Coenen and Vendrik, 1972; Burke & Cole, 1978). The most common factor that are difficult to two-fold or varies is the animal's level of arousal which can shift often in several minutes. Any difference found between a binocular - 5 response tested over a period of a few minutes and a monocular response taken a few minutes later could be due simply to changes in neuron responsivity (Figure 1). In this work, a new strategy was used to study the binocular interactions in the LGN. First, control is exerted over parameters of visual stimuli that are known to profoundly influence the cortical cells. Moving slits, edges, and orientation, direction, displacement. interactions relative were retinal studied by comparing of gratings were used and tested with systematic changes in and firing sweep velocity Second, binocular and and binocular monocular responses taken in tandem. Computer operated shutters placed over the animal's eyes allowed continuous alternation binocular tests. Histograms were between constructed monocular and from the difference between each monocular and binocular paired sweep as recorded by computer. Third, complete statistical analysis was applied to all his- tograms so constructed to clearly identify real binocular interac- tions. We have found that the great majority of binocular inhibition and facilitation amplitude of the binocular interaction stimulus and is LGN cells that the polarity and strongly function both of velocity, position, retinal disparity, and sweep direction. The functional significance of these interactions that a show and the evidence they are due to influences from the cortico-thalamic pathway is discussed. - 6 - Figure 1 This figure shows the responses of two cells each recorded over a period of about 28 minutes to stimuli sweeping over their dominant receptive field (RF). The top cell LA7-1A (CONTRA-Off) was stimulated with a slit and the bottom LA7-4 (CONTRA-On) with a 1.7° grating at three different velocities. The left column for each shows a histogram for the first 14 minutes (50 trials) and the right column for the next 14 minutes. Notice that the amplitude of the response peaks has changed considerably. In comparipg LGN monocular responses taken at one time with binocular responses taken at another, one cannot expect to get a good measure of the difference. Taking both binocular and monocular sweeps in alternating succession is the strategy used in this study. - 7 - LGN RESPONSE VARIABILITY FIRST 50 TRIALS I NEXT 50 TRIALS f3oom• I CELL LA7-1A 41••• / CELL LA4-4 Velocity Tune 85 •.• / 11110Figure 1 - 8 2. METHODS 2.1 Electrophysiology Nine normal adult cats were prepared for single unit recording using standard methods (Barlow et al., 1967). Anaesthesia was induced with 4% halothane and then shifted to 0.5% to 2% nitrous oxide and oxygen. in a 2:1 mixture After insertion of venous and tracheal cannulae (and in some cases a bilateral sympathectomy to movements), the animal eye paralytic mixture to minimize eye movements (Rodieck et al., 1967). The mix- ture delivered 5 mg/kg/hr Flaxedil, 0.5 mg/kg/hr 0.1 reduce was transferred to a stereotaxic headholder and put on a continuous intravenous infusion of a designed of d-tubocurarine and to 0.4 mg/hr Dexamethasone in 5% dextrose and 0.25% saline at 5 ml/hr. The animal was artificially respired at 40 breaths/min with a 75% nitrous oxide, 22.5% oxygen, 2.5% carbon dioxide mixture at 30% hyperventilation relative to the Harvard respirator Heart rate was monitored. recommendations. A rectal thermometer and electric blanket circuit were used to keep body temperature at 37.5°C to 38°c. The pupils were dilated retracted with neosynephrine, power contact lenses. tangent with Cyclogyl, and nictitating membranes the corneas protected with zero An image of the retina was projected on a screen 57 cm away using the fiber optic technique (Pettigrew et al., 1979), and the positions of the optic disks (and trales when visible) were plotted. cen- When not visible, the areae cen- trales (AC) were taken to lie 15.6° nasal optic disks (Nikara et al., 1968). areae and 6.8° down from the - 9 Tungsten-in-glass electrodes (Levick, 1972) were lowered toward the LGN through a craniotomy located on or near 6 mm anterior and 8.5 mm lateral to Horsley-Clark zero; variations were used to sample different parts of the visual field. For some animals, a second elec- trode was placed into area 17 through another craniotomy near the midline of the skull at 1 mm posterior to AP zero. Cortical electrode guides were held in a chamber filled with agar and sealed with paraffin to maintain recording stability. LGN recordings were found to be stable without a chamber. the Most often, electrodes were advanced brain with a custom-built stepping motor microdrive system (Cen- tral Engineering Services, Caltech). positioning This system into enabled of electrode depth with minimum delay. electrode were picked up by a high input impedance fed into precise Signals from the preamplifier and an oscilloscope, audio amplifier, and spike discriminator. Output from the discriminator was fed into a Nova computer (also used for stimulus control and data analysis - see below). Recording ses- sions lasted up to 48 hours. The LGN electrode was lowered to 11 mm from the surface automatically over a period of from 0.5 to 1 hour. It was then lowered under hand control until strong modulation of background activity stroboscopic illumination and by shifting of a hand held grating (at all orientations) was obtained. Single capable of triggering the units with isolated spikes discriminator were then sought. An LGN unit was characterized by a clear center-surround RF with a crisp or Off by On response to a spot flashed in the RF center in only one eye. LGN tracts always started with purely CONTRA cells and then showed a - 10 - clear transition to IPSI cells at a greater depth, followed again by CONTRA cells even deeper. - 11 2.2 Visual Stimulation and Protocol A visual stimulus was projected onto small mirrors puter controlled mounted on com- pen motors capable of placing an image anywhere on the tangent screen. The stimuli were most often slits of adjustable dimensions, but gratings on 35 mm slides were also used. The orientation of a stimulus could be controlled manually or by the computer. A plotting table received an image of the tangent screen reflected from a plexiglas sheet mounted appropriately between the screen and the grating and projector. Typically, an LGN unit was found with a hand then held plotted in detail using a 0.25° spot controlled manually with a joystick and an on/off switch. each unit before The following data were recorded binocular interactions were studied: 1) electrode depth, 2) eye dominance (CONTRA or type (On or for Off), 4) RF IPSI), 3) receptive field (RF) configuration, including diameter, boun- daries, and annulus extent, and 5) position relative to the dominant eye's AC. The study of binocular interactions was started for each cell by determining and correcting for the vergence of the two eyes. The vergence was estimated in one or more always employed and involved the of three The first was above described plotting of the positions of the ACs. The second was used when recorded ways. cortical units were (in about half of the animals); the average position of the RFs for each eye was determined from several strongly binocular cortical neurons. This yielded a physiologically accurate measure of - 12 vergence. The third could be used only after crossing a laminar border in the LGN; vergence was taken as the shift in the position of two units' RFs when the units were driven thus from different eyes and were on opposite sides of the border. The vergence estimated by these techniques was eliminated by placing a prism in front of one of the eyes to shift its AC to the same position as the other on the tangent screen. The apparatus used to study binocular interactions, as all each eye, Figure 2 shows two 5 mm artificial each of which trolled both pupils, The shutters by hand and from the computer. able prism over the right eye that is driven could be con- Figure 3 shows a variby a stepping motor, controlled by hand or from the computer. This prism was used to vary binocular displacement and could be placed over either also eye. It provided a mounting for the vergence correcting prism described above. Over the left eye is a dove prism which was occasionally to one could be closed independently by a shutter driven a by small D.C. motor. also as of the equipment physically in contact with the animal, is shown in Figures 2 and 3. over well present different sweep directions to used each eye. The variable prism, dove prism and other optical devices could be mounted in front of the shutters (as in the figure). A small laser was used to insure proper alignment of all optical devices. The reflective tapetum was visible through all optics when viewed through an ophthalmoscope held at the position of the tangent screen 57 cm away. - 13 - Figure 2 This figure shows the preparation from a the point of view of the tangent screen, with all correcting and modifying prisms moved to the side to show the artificial pupils and shutters. These shutters could be opened and closed independently under manual or computer control. Other visible items include 1) electrode microdrive pushing electrode into the brain, 2) preamplifier, 3) computer drivable variable prism (pulled aside), 4) tracheal tube with hoses leading to respirator, 5) cat, 6) artificial pupils with motor driven shutters, 7) dove prism (pulled aside) and 8) aluminum foil shielding to reduce electrical noise. Figure 3 This is similar to Figure 2 but with the stepping motor variable prism and dove prism put in place. The variable prism also acted as a mounting for the vergence correcting prisms. Notice that the shutters and artificial pupils are still in place. - 14 - Figure 2 - 15 - Figure 3 - 16 The NOVA computer was programmed to control the described apparatus and the visual stimulator in a coordinated fashion so as to study •binocular i nteractions. In general, the user first constructed a "run" by describing to the cont r olling program the set of values to be tested for a particular visual parameter. In the the present study, orientation, direction, velocity or relative binocular displace- ment could be varied during a run. All parameters held constant ing a dur- run, such as slit size, illumination level, non-slit stimulus (on a slide), sweep duration, sweep center position, number of trials and so on could be reset for each run. Second, the user defined the shutter settings to computer was as used. It then runs in the following an example, for a velocity tune. One of the test veloci- ties is chosen pseudo-randomly and the shutters are instantly for the randomly). of The asked to test the cell response for the right eye only and/or left eye only and/or both. manner, be right eye, the left eye, or opened both (also chosen pseudo- The stimulus is then swept across the screen and the time occurrence of each spike with reference to the sweep is recorded by the computer. The shutters are then reset to one of the positions not and yet tested the stimulus swept again at the same velocity, continuing until all of the shutter positions are run. The process is then repeated with the next velocity. The entire set of velocities are tested in this manner for a preset number of trials. The LGN shows large variations in response amplitudes over a few minutes (INTRODUCTION). just With the above interleaving technique, - 17 the separation in time between each monocular and binocular sweep was never more than a few seconds. Accurate measurement of the differ- ences between monocular and binocular responses could independent then be made of shifts in a cell's overall responsiveness. It became possible to construct "binocular difference histograms" by subtract- ing each dominant-eye sweep from the binocular sweep taken in tandem and then adding all of these "difference sweeps" together. The general protocol was to find a cell, record all (described data earlier), eliminate vergence, and then to study binocular interactions as direction basic a function of velocity, binocular displacement, of sweep, and other parameters as time allowed. Most often horizontally moving vertical slits, edges and gratings were used. The computer displayed a wide-bin histogram for each shutter setting as the experiment was running and a crude visual estimate of the binocular differences were made. In general, two to four sets of three to seven velocities ranging from O.5°/sec to 4O°/sec were run first an and estimate made from the display as to the velocity producing maxi- mal binocular interaction. That velocity was then used in a series of binocular displacement tests (using the computer driven variable prism), usually run at 1° increments over of range of 8°. ments appearing to show Displace- greater binocular interactions were often then studied with finer test increments. Occasionally, binocular displacement tunes were run at more than one velocity. Following this, direction tunes comparing responses of vertical stimuli moving toward versus away from the AC were done at various velocities and displacements. Occasionally full velocity tunes and/or displacement tunes - 18 were done for both directions. At times, full orientation tunes were run. Other tests were run as time allowed. Cells were rarely held for more .than five hours - enough time for all of the above tests and more. More often, a cell would be lost before p le ted. all tests were com- The described protocol was varied often to allow completion of a significant number of some of the latter tests. - 19 2.3 Data Analysis On the average, 40 runs of from 10 to 100 trials each over a range of from 2 to 10 parameter settings were obtained from one cat, with an average of five well tested cells per cat. This often resulted more than 1.5 megabytes of information recorded per cat. in The goal of the analysis was to produce clear, accurate, and reliable measures of monocular responses and binocular interactions from these data. analysis involved putting the data series file from each run The through a of programs which 1) extracted monocular histograms and bino- cular difference histograms (BDHs), 2) checked each BDH for intervals of statistically significant differences (between the monocular and binocular response), 3) extracted amplitudes of inhibition or facilitation for each of these intervals, and 4) displayed all of the data and analysis results in a variety of usable charts and plots. Each of these steps is de f ined more clearly below. The results were then collected together by hand to build up the final tuning curves and represented in the RESULTS section. maps It should be noted that simula- tions of a neuron with preset monocular responses, binocular interactions and statistical variability were programmed to create runs to test and debug these steps. The analysis programs were shown to accurately extract the data that had been set into the simulation. BDHs were built up "moment" (in units for each sweep moment then subtracting for each of 3 msecs for technical reasons) the number of spikes in the dominant eye sweep from Each by its paired binocular sweep. had from -3 to 3 spikes as the difference between - 20 the binocular and monocular response and a sweep typically of 300 to more than 1000 such moments. adding moment by moment all of the consisted A BDH could be built up by sweeps so generated. If four velocities were run, four BDHs would be built. The binocular difference occurring during a time interval BDH was calculated The computer scanned a BDH and for t-values for intervals of when one moment was together as Overlapping intervals were one interval of real binocular difference. The maximum spikes/sec difference found in small moment significance exceeded 99.5% on a double-sided test (most demanding criteria conventionally used). assigned every a large range of sizes. The interval size producing the maximum t-value for any joined a tested for statistical significance using the paired t-test (Duckworth, 1968). noted on as its enough amplitude. any twelve msecs interval was Twelve msecs was chosen because it was not to dilute sharp firing and technically convenient. representative rate peaks, large enough to produce a Inhibition was represented by negative amplitudes and facilitation by positive. A variety of plots and charts were produced from examination and for construction of tuning each run for curves: 1) monocular response histograms, 2) statistically filtered and unfiltered BDHs at various boxcar sizes (Figure 4 explains boxcar histograms and Figure 5 shows filtering), and 3) printed charts of all exceeding intervals near or statistical threshold, including for each at-value, dura- tion, amplitude, corresponding monocular amplitude, etc. The various tunes and maps described in RESULTS were then put together by hand by combining results from the appropriate runs. - 21 - Figure 4 This figure shows the use of the boxcar technique to visualize binocular difference histograms (BDHs). Normally histograms consist of neighboring bins that are much wider than the minimum measurement "grain", which is 1 msec for spikes. All monocular histograms in this paper are so built. If a peak response occurs over a duration less than or equal to the bin width, it does not show fully if it is out of phase with the bin construction, which is usually arbitrary. Boxcar histograms are constructed using bins of constant size built at every grain, as if each bin were a boxcar travelling along the time axis. Phase is never a variabl~ in this type of histogram; the only variable is the width of the boxcar. Three widths are shown here. Each width is analagous to a band pass filter and, for example, boxcar histograms with narrow bins will optimally show short duration responses. A range of narrow and wide boxcar BDHs were prepared to scan binocular differences because no good information predated this study as to the timing of binocular difference responses in the LGN. Note that the amplitude decreases with larger boxcars because of the dampening that results from averaging peaks with neighboring lower response frequencies. - 22 - BOXCAR HISTOGRAMS Binocular Differences ff lOm1 BOXCAR Monocular Response: Cell LA 2 -s I 200 ffll I H 90m1 BOXCAR .•"' 0 -~ "' 4s' C: 1 _ _ _ _ _...JlllL.._ -----=i.afl....__ _j~ .:: 360mo BOXCA Figure 4 - 23 - Figure 5 This figure shows the monocular response and BDHs for a partial displacement tune. In general, BDHs were prepared for viewing as collected from the cell. After t-values were calculated for all intervals along the BDHs, all parts that did not meet a chosen statistical threshold were eliminated, and parts that were near the threshold were scaled down. In this case, three peaks of similar amplitude show different levels of significance and thus the middle one is reduced by filtration at the 99.2% level. For the purpose of scanning BDHs, filtration was done to that level, but for the purposes of data presentation in th i s thesis, filtration is usually done to the 99.5% level, the chosen cutoff for defining a binocular difference as real. - 24 - STATISTICAL FILTERING: CELL LA2-5 MONOCULAR RESPONSE ~ - 1.5° _ I~ .............._-f"--_.__-.L.I.L..___ _- + - - - z w :I w u ..,C Q. Ill c .., a: 0 .5 o -~~-r-.......___.................___ _--+-'...___ C ::> u 0 z ID BINOCULAR DIFFERENCE HISTOGRAMS UNFILTERED FILTERED TO p < .008 I Figure 5 - 25 3. RESULTS 3.1 General Si xty-one LGN cells were plo tt ed from a series of Twenty-one (34%) nine adult cats. we re CONTRA-On, twelve (20%) were CONTRA-Off, six- teen ( 26% ) were IPSI-On, nine ( 15%) were I PSI-Of f , CONTRA-Unclassi f ied, was IPSI-Unclassified. and one ( 2%) t wo Binocular Difference Histograms (BDHs) were obtained f r om ( 3%) were Usable forty-four of these units. Binocular Feedback (BF) 1 was found in the majorit y of LGN cells. Each unit was tested for BF varying as many different parameters of the visual s timuli a s were poss ible unt il its spike was lost. BDH was comp uted for each parameter, filtered to p < 0.005, and A a firing frequency for each interval of inhibition and facilita t ion was calculated as described in METHODS. Thirty-seven (84%) ce lls showed significant binocular faci lita t ion and f orty-thre e ( 98%) had significant binocular inhibition for some visual stimuli. stimulus parameters (velocity , direction, etc.) The number of th at c ould be tested varied fr om cell to cell and one might e xpect that the probability of finding BF woul d be less f or cell s gi ven fewer tests. When cells test ed for less t han three computer runs are excluded, a larger centage per- sh ow clear BF, with twenty-six (93% of 28 ) showing facilita- tion a nd twenty-eight (100% of 28) showing inhibition. 1. LGN cells receive only monocular retinal input, and so all binocular interactions are referred to as "feedback" without necessarily presuming the source (e.g., interlaminar, cortico-geniculate, etc). - 26 Maximum facilitation and inhibition can be compared to cal 2 monocular response firing frequency for each The cell a typi- (Figure 6). t_ypical firing frequency for all cells ranged from 40 spikes/ sec (s/s) to 247 s/s, with an average at 119 s/s. cular facilitation was 52 s/s (0 to 156 s/s) and the mean maximum binocular inhibition was 71 s/s (0 to 147 most likely an The mean maximum bino- underestimate of These s/s). the actual values. values are Removing the cells tested for less than three runs increases the facilitation mean to 62 s/s and the inhibition mean to 81 s/s. No significant correla- tion between monocular response and seen. amplitudes BF amplitudes was Twenty-one (48% of 44) cells showed facilitation exceeding 50% of their typical firing frequencies and six (14%) exceeded 85% of the typical monocular response. Similarly, twenty-seven (61%) had bino- cular inhibition to a level 50% below the typical monocular level and eleven (25%) below the 85% level. Clearly, the majority of the LGN cells tested showed clear and strong binocular facilitation and inh ibition with appropriate visual stimuli. The occurrence and magnitude of BF was not significantly dif- ferent for CONTRA versus IPSI cells or for On versus Off cells (Fig- ure 6). Cells with more than three computer runs and with receptive fields estimated to be within five degrees of the Area Centralis (AC) averaged 43 s/s for maximum binocular facilitation while similarly run cells with fields more than five degrees away averaged 80 s/s for facilitation. cular No difference was found between these groups for bino- inhibition. BF was therefore plotted against estimated RF 2. "Typical" is defined as the median of the peak monocular firing frequencies measured in all tests. - 27 - Figure 6 Maximum significant binocular facilitation and inhibition is plotted for each cell against its typical monocular firing frequency. "Typical" is defined as the median of the peak responses found in all runs for that cell. No significant difference in the distribution of Binocular Feedback (BF) is found between CONTRA and IPSI or between On and Off cells. No clear correlation is seen between typic a l firi n g frequency and BF amplitude. Figure 7 Maximum significant binocular facilitation and inhibition for each cell is plotted against the estimated distance of its monocular receptive field (RF) to the Area Centralis (AC). Only cells for which three or more computer runs were taken are included. No significant correlation is seen. The triangle plot-points are facilitation and the circles are inhibition. - 28 - DISTRIB UTION OF MAXIMUM BINOCULAR FEEDBACK ONTO ' CELLS 150 0 10 Fe c 1 11t et 1 on s p 1 50 k 0 s D e I 0 s C 0 n 50 Cl 100 • • • • • I D D aR D •• 0 lo •• I 0 I 0 r. • g i D • D I • •.I! •I I Inh 1b1t1on D D • D e .• ... •• ~ . I - CONTRA ON D - CONTRA Off - JPS! ON 0 - IPSI OFF I • 150 0 ~ 0 0 0 I 0 I • •• I I 0 I 150 Typ1ce1 Monocular Peak Response (sp1kes/sec) Figure 6 D • - 29 - DISTRIBUTION OF BINOCULAR FEEDBACK AGA INST RECEPTIVE FIELD DISTANCE TO AREA CENTRAL.IS t, 15 0 t, Fa c1 11t a t1on 100 s p t k t, t, t, t, t, t::. t, t::. t::. 50 i e t::. t::. t::. s I s e C 0 n t, t, 0 5°[ 0 d 0 0 0 0 - 1 OJ Jnh1b 1t1on 0 0 0 0 0 8 -,J ~15 10 of § 0 0 0 0 0 0 0 0 § 0 Dfs t a nc e f r om Are a Ce n t r a lf s ( de grees ) Figur e 7 - 30 distance from the Area Centralis (Figure 7). statistically Howev er, no clear or signific ant correlation is found acro ss these measures and no difference in facilitation between central and peripheral ce lls is reliably demonst rated . It i s most likely that these measure s of t he oc currence and maximum amplitude of BF in LGN ce lls are i n general l e ss than the actual values. not Differences between binocular and monocular taken as BF unless significance exceeded responses were 99 . 5% on the more demanding double-sided paired t test. Furthe r , as will be described in later sections, BF is often well tuned to variations in parameters of visual st imuli . Reliably measured maximum BF must always be equal to or less than the maximum obtainable BF with totally ideal stimuli. Given the limited time available in testing for BF, the idea that the above measures are underestimat es is support ed . The large maj ority of LGN cell s tested show strong and cant BF under at least some conditions of visual stimulation. The ampli tude of the maximum BF to be found for each cell conditions is signifi- largely independen t of the size under of its monocular responses, and for half or more of t he cells exceeds a typical cular response by at least 50%. optimal mono- Binocular facilitation occurred only slightly less often than inhibition, and the averag e amplitude of facilitation was just slightly less than that of inhibition. The frequency of occurrence and the amplitudes of maximum BF were not to found be related to IPSI versus CONTRA or On versus Off categorization. Cells further from the AC showed an apparent increase in facilitation - 31 - amplitude, but this was not statistically significant. - 32 - 3.2 Location of Binocular Feedback in Visual Space Typically a unit was studied with a long vertical izontally mov ing binocular displacements (see test BF METHODS). erated in one run for a range from four at di f ferent relative Most often, BDHs wer e gendeg r ees converged to fou r diverged, sampled at one degree intervals. Further runs were often done t o expand the range or to focus in on a test intervals. particular with finer were chosen either on the bas is of other earlier tunes, if BF hor- across it s RF . A computer driven variable prism was placed in front of one of the two eyes t o degrees slit rang e Other parame ters of the visual stimuli any, of fo r that cell (e.g., using the velocity that yielded the greatest BF) or on the response. basis This of producing a clear a nd strong monocular technique revealed many regions of BF that are well tuned for r elativ e retinal di sp arity. A composite map of the locations of BF i n v isual space was duced fo r each cell t e sted (Figure 8). maps The meaning of BF pro- location can be made clea rer by considering how the ocular location of a region of BF would r eflect on t his type of map (Figu re 9). A region of BF with a specific location only on the dominant retina shows as a ho r izont a l band if the variable prism is plac e d over the non-dominant eye and as a diagonal band when the prism is over the dominant eye. Similarly, BF localized only on the non-dominant retina diagonal band as a in the former case and horizontal in the latter case. If the BF is found only over a limited range of ments, shows suggesting binocular displace- a requirement for the binocular retinal disparity - 33 - Figure 8 Ma.ps of b inocula r feedbac k are created in a series of stages. Cell LA3-8A (CONTRA- On, RF 4° from AC) was tested with a vertical slit moving ho rizontally at 8°/sec towards th e AC. The variable prism was placed over the non-dominant (IPSI) eye and 9 dif fe rent ver gences were tes t ed . The monocular resp onses at e ach setting are similar. The BDHs show a clear region of binocular inhibition that is approximately 2° wide (horizontally) on the retina but that can be elicited only in a sub-range of binocular displacements; fro m -1° to 3°, clearly peaking at 1°. The BDHs are filtered to p < 0.005 to identif y re gions of BF, and the regions are map ped graphically to demonstrate t heir location in visua l space. The abscissa ("Relative Retinal Pnsition") refers only to the dominant e ye, as the posit ion on the nondominant eye shi fts by 1° for each 1° of displacement. The dotted line represents the relative location of the unit's monoc ular receptive field. The amplitudes of responses seen in BDHs will in general be less than in the maps because BDHs are constructed with large boxcars (see METHODS). In this paper, negative displacements are converg ent for CONTRA cells and d ivergent for IPSI cells . In t his and all figures, "s/s" is an abbreviation for "spikes per secon d". - 34 - Monocular Response for Cell LA3-8A Binocular Difference Histograms UNFILTERED FILTERED To p<.oos 0.8° I ... 100 ma I s 4• C r-- <l.) E (l.) 3• u N M !! M" r-- ('O Ci. 2· Vl Cl ,_ 1· ...___,,__c._ -+-<=.__ _..,._ _ _ ('O ::i o· ..,_=ll--_..c.--=,____---".___--t-CJ.~ ~ C co Q..) ... -1· > ..:::: -2° ~ . - j l . , / J - - " - - - - " - ' - - ~ - Q..) Ct: INHIBITION r-i p ? .O0S t=j to 36 • •• .,_J Binocular Feedback Map r C o □ 37 - 72• • P < .00S [ [ ] 73 - 107 .,. 0.5° o· ,,,..... ······· ·· ··· ···· ······· ···· ··· ······ ··••·• ··· ····~ ···· ,L__ \-\ \-- ,r--- ---- '\--- ... - 1° -- -, / / / / '--- ('O <l.) Ct: ~ib-~ -1.5 .........,....-.....- - - . - - - - , - - -.....- - . - - - - - - . , - - -.....- -....._ . -2· -1 · o· 1· 2· 0 <J') Relat ive Binocu lar Di splacement Figu r e 8 - 35 - Figure 9 BF location maps are more clearly understood in terms of ocular location. The left-most column represents the position of idealized regions on the dominant and non-dominant retinas in relation to a slit sweep at various binocular displacements. The center column represents idealized BDHs for three arrangements of BF and the third column represents the resulting BF location maps. This figure assumes the variable prism is placed over the non-dominant eye. Region on Dominant Retina: BF driven from a specific region on the dominant retina (and with no specific location on the other retina) would produce the same BDH no matter what the binocular displacement. Maps of this type of region would contain a horizontal band, the thickness of which would represent the width of the region. If the prism is placed over the dominant eye, the band would be diagonal. Region~ Non-Dominant Retina: BF driven from a region on the non-dominant retina (e.g., inhibition from the other LGN lamina) would produce BDHs differing only in the time of occurrence of the peak difference. The map would contain a diagonal band (or horizontal, if the prism is over the dominant eye). Disparity Tuned Region: BF requiring specific binocular displacements (or disparities) would produce peaks on only some BDHs. The map would contain "islands" representing the BF: its location on the retina shown by its abscissa value and its location in depth-space shown by its ordinate value. - 36 - MAPPING AS A FUNCTION OF OCULAR LOCATION Region on Dominant Retina --------_ _,.1111...___ .. ~r------'---------►~--Slit Sweep . . . ._ - - - - ~ ~Region on Non-Dominant Retina 0~ / ~ -,.~ _____,,,,A_ I 2 I/ iu 1o11 • ~ / -1 iia. _ _ _. , . _ . . __ _ _ mapped _.A_____ .A.__ I ~~ I - I Disparity Tuned Regi on 2 Region on Non-Dominant Rec ina Region on Dominant Retina -lo, _ _ _...,..ai_ _ _ __ -20_ _ _ _ _ _ _ _ __ ~o ·Qo Di spla cement Figure 9 lo - 37 of the stimulus, an "island" would show on the map. Figure 8 contains a clear example of such an island. Further, in all cases, the width of the region on the retina can be determined from the width of the band or island along the abscissa of the map. Complete BF location maps were made for twenty-five cells vertical slits or edges using moving horizontally across the retina. In many cases, maps were made at different velocities, sweep directions, etc., and maps from the same cell under differing stimuli parameters were often quite different. A total of thirty-nine Every maps were made. cell and every map showed considerable BF. Further, the major- ity had multiple regions of BF with both facilitation and occurring at different locations. from a unit with a clear facilitory island. This cell also shows another presence of strong BF at Figure 10 shows histograms made island next feature to an commonly inhibitory found: the the onset and at the offset of the slit sweep, even though the slit comes on and goes off at up from inhibition to 8° away the monocular receptive field. This onset and offset phenomenon will be described more fully later. Of the twenty-five cells, only one had a map of pure (Figure 8) and only one had a map of pure facilitation (containing two distinct islands). A more island of BF surrounded polarity (Figure 11). region of BF inhibition common arrangement consisted of an on the location map by BF of the opposite A similar mapping seen consisted of a large with a smaller region of BF of the opposite polarity, where the smaller region ( a "peninsula") is not contained by the - 38 - Figure 10 This figure shows one of the computer runs used in making the BF location maps for LGN Cell LA.9-10 (IPSI-O·ff , RF ll o from AC). The computer driven variabl e prism was placed over the dominant eye in this case and as a result the monocular response shifts by 1° for each degree of binocular displacement. The dotted line on the BDHs represents the position of the RF as determined by the monocular response peaks. Clear neighboring islands of inhibition (peaking at -2° to -3° displacement) and facilitation (peaking at +3° displacement) are seen, each located differently on the retina and in relative depth space. Also clear is the presence of sweep onset and offset binocular inhibition. (A complete map of BF from this cell is shown in Figure 32). - 39 - Binocular Feedback Mapping: Cell LA9-10 Monocular Response I 300 m. 2.4 degs I ...... C Cl) E -30-~1--A~~------....c.=~-~j-L-I.-. Cl) u C<:l Cl.. -4°_--.JUL--~-------....t!!:.~~--..Ju._i.z::.- IJl 0 ,_ Binocular Difference Histograms ro :J u 0 Facilitory Area C ! O'.l Cl) 40 w Cl) 1300 ms 2.4degs -----t.r""IW. .........,~----..-....__,.........._..._.,,_.,....,.._, > ·...... 0:::: OTet __.I~ 2° _ ..........r--.,.....,....-___,~_.,....._.._.....,.,........,------t..r--.--." Figure 10 I - 40 - Figure 11 This figure shows the binocular feedback map for LGN cell LAS-6 (IPSI-Off, RF 4o from AC) taken with a vertical slit moving horizontally towards the AC at 12O/sec. A clear island of binocular facilitation is seen to be imbedded in a larger region containing binocular inhibition. The requirement for both binocular displacement and retinal position is sharper in this cell for facilitation than inhibition. The dotted line represents the position of the monocular RF of the cell. The variable prism was placed in front of the non-dominant eye for this test. The BF seen at sweep onset and offset is also depicted. Figure 12 This figure shows the binocular feedback map for cell LA9-7 (CONTRA-On, RF 14° from AC) taken with a vertical slit moving horizontally away from the AC at 8°/sec. A large region of binocular facilitation is seen next to a smaller region of inhibition. Tests were not done at displacements more than 4° divergent for this cell and it is only clear that inhibition at least requires displacements greater than or equal to 3° divergent. The dotted line represents the position of the monocular RF of the cell. The variable prism was placed over the dominant eye for this test. No BF was found at sweep onset. The offset BF consisted of facilitation followed by inhibition, as indicated by the triangular splitting on the map. - 41 - MAP OF BINOCULAR FEEDBACK FOR CELL LAS-6 ,ACILITATION :1 1,: . • P~-005 INHIBITION c:J (:J P<.005 ■ 15-29•• ~ 34-66M ON , ......... '' I - I' 0~ 00 ...,, ci: 00---= = ~ - .::: ~ ~ 0 ~-2 0 -4 OFF -20 -10 00 10 Relative Binocular Displacement Figure 11 TO 33 ~• ~ ■ 30-42M r --t-:: p~.005 20 67-100 ~- - 42 - MAP OF BINOCULAR FEEDBACK FOR CELL LA9-7 FACILITATION D ,--.I INHl ■ITION p? .00!5 L -- p< .005 OFF __ 6" V\ 0 ~ -2· ~ fb !Al ClJ > ( 'tl ! Cl) 0:: -4· -6· .....- ...- -....- -....- -....- - ~ . - - - - . . - - -.....- -....- -....- - 2· Relative Binocular Displacement Figure 12 - 43 larger region and the smaller region occurs on one of the extreme displacements tested (Figure 12). Five cells showed a of facilitation surrounded single island inhibition. An additional two cells by showed a single facilitation peninsula and larger regions of tion. inhibi- A single inhibition island surrounded by facilitation was seen in three cells and two cells showed inhibitory peninsulas next to extensive facilitation. Two cells showed a single island of facilitation paired with a single arrangement, however, island of inhibition. maps taken most common consisted of two or more regions of facilita- tion and two or more regions of inhibition. teen The from This was seen in seven- twelve cells. BF clearly can be found, it is seen, in a number of different arrangements. Every map from every cell contained at region at a specific retinal binocular displacements. Nine least one a island location and at a range of specific cells also showed bands of BF as described above, demonstrating the existence of BF regions which have a specific location on the retina but which lack a strong requirement for binocular displacement (Figure 13). gle Four of the cells had a sin- band of facilitation, three had a single band of inhibition, and two had one of each. As described above, the orientation of the (horizontal versus diagonal) can be region has a well defined location on the retina. used to indicate whether the dominant or non-dominant Four of the cells had a band indicating the dominant retina, four indicating the non-dominant, and one had a band Figure band 32). BF amplitude for each (see It should be noted that even within bands, increases in are often seen at particular displacements, which - 44 - Figure 13 This figure shows a binocular feedback map for LGN cell LA2-3 (CONTRA-On , RF 12° from AC) taken with a v ertical slit moving horizontally away from the AC at 12°/sec. Three strong regions of BF are indicated, two facilitory and one inhibitory. The central inhibitory region spans all displacements tested and forms a band parallel to the monocular RF (dotted line), thus indicating a weaker requirement for binocular displacement than the cells depicted in the two previous figures. The tests were run with the variable prism over the non-dominant eye. - 45 - MAP OF BINOCULAR FOR CELL FEEDBACK LA2- 3 FACILITATION D ~ ~ P< .005 INHIBITION p ~ .oos[_-J r-=; TO 16 1 1 • ~ ■ 27-53 1 1 8 ~ ----- ■ 54-80 818 1;;;;-... ~ "i ....C: Q. w w ~ .s ~ ! s - -=..:--"""_'-_-_----·. ............. C: .2 .... 0° Vl ~ <lJ > 0 ·:: -2 ~~~ ~ & \ ON Relative Binocular Displacement Figure 13 - 46 indicates some preference for a particular retinal disparity. Inhibition and facilitation were fairly evenly distributed among the population of cells tested. Twelve maps from ten cells could be classified as predominantly containing inhibition, fifteen maps from twelve cells containing mostly facilitation, and the remaining twelve maps from eleven cells had roughly equal amounts of facilitation inhibition. and The relative predominance of one polarity of BF could in fact switch in the same cell depending upon the visual stimulus particular arrange- presented (see Figure 32). All of the maps were checked to see if any ment of BF could be correlated with other factors. No significant difference was found in BF arrangement, predominance of inhibition or facilitation, or the occurrence of banding for IPSI versus CONTRA, On versus Off, or central versus peripheral cells. Although most regions of binocular inhibition are and facilitation similar in that they both have specific requirements for binocu- lar displacement, they are quite different in their distribution vergence space (Figure 14). of (15 of 51; 29%) found The number of at islands in facilitation an estimated zero degrees displacement almost doubled the number of inhibitory islands found there (8 of 54; 15%). Special precautions were taken to get an accurate measure of the vergence of the eyes, often including the use of reference cortical units (METHODS). Although the relative vergence of the retinas of a paralyzed cat can be hard to measure, and is believed to have slow a drift of less than one degree in an experiment, the tendency of - 47 - Figure 14 51 island s of facilitation and 54 islands of inhibition from 39 maps f r om 25 cells were plotted in the bin representing each one ' s optimal binocular displacement. In this figure, negative displacements are convergent. Although errors in measuring actual retinal displacements would ten d to cause a "spreading out" of both histograms, the increased tendency of facilitory islands to lie at zero degrees displacement is clear. There is a also a tendency for islands of both facilitation and inhibition to require convergent displacements (see Text). So-called "peninsulas" have been excluded from this figure, due to the uncertainty of their actual preferred displacements. - 48 - Binocular Feedback Locations: Optimal Binocular Displacements 15 FACILITATION 12 9 6 3 • C 0 al • a: -.. -50 -J.o 0 -30 -t> -10 00 10 2° 30 40 50 -10 00 10 20 30 40 50 • SJ E :, INHIBITION z 9 6 3 S-5° -40 -30 -2° Optimal Binocular Displacement Converged Diverged Figure 14 - 49 - such errors would be to flatten out the histograms in Figure 14 and to do so equally to facilitory and inhibitory regions. The difference between facilitation and inhibition is even more striking in light of this and it is likely that the data represent an underestimate of the number of r egions that actually lie at zero degrees displacement. BF also showe d a t e ndency to be convergent. Altho ugh "peninsu- las" are not included in Figure 14 because their optimal displacement is not clear, the vergence of each peninsula is clear. Including them gives 65 convergent regions and 56 divergent regions. Inhibition and facilitat i on are not strongly different in this tendency. It should also be noted that the 40 peninsula regions not l nc1uded in Figure 14 all exist at the more extreme displacements, as described above, BF was found even as extreme as six degrees diverged and eleven degrees converged. It is unlikely that this is totally due in measuring retinal and to error displacement as in three cats four successive cells on the same electrode track had BF located over large ranges of displacements (Table 1). - 50 - Layers Crossed Cat !RF Distance! Range of Displacements Found: to AC I Facilitation I Inhibition I I (degs) I (degs, C=converged, D=diverged) ------------------------------------------------------------Al 15 >4C to 2.SD SC to >6D 1 LAl LA2 LA7 A to Al A to Al I 16 to 6 I 4 llC to 4D SC to 6D 4C to 4D 7C to 4.SD Ranges of Optimal Binocular Displacements in Single Tracks of Four Units Each Table 1 Comparisons of optimal vergence categories. No were made between different cell significant differences were found between IPSI and CONTRA, On and Off, or peripheral and central. Regions of BF can also be described in terms location relative of retinal to the location of their monocular RF on the dom- inant eye. The relationship is depicted in Figure 15. of lies regions their in The majority the same place as the monocular RF, although a number do not. The mean distance for all regions is 0.2° with a standard deviation of 2.4° indicating fair scatter. These measures are considered as very accurate because the position of a and the monocular RF are region of BF determined simultaneously from the same sweeps of the visual stimulus. No difference is found between inhibitory or facilitory BF. Figure 16 shows the size distibution of regions of BF. Horizon- tal width ranged from under 0.1° to as large 5.8°. Facilitation averages 1.4° and inhibition averages 2.0°. These measures are most - 51 - Figure 15 This figure shows the distribution of the distances of the center of a region of BF on the dominant retina from the center of its unit's monocular RF. All regions of BF that could be located on the dominant retina from their BF maps are included in this histogram. Because all maps were made with vertical slits moving horizontally, the distances depicted are horizontal. Negative distances are on the AC side of the monocular RF. The mean distance for all regions is 0.2° with a standard deviation of 2.4°. There is no difference between facilitation and inhibition. Figure 16 This figure shows the distribution of sizes of regions of BF on the retina. Size is calculated as the region over which the binocular difference is reliable beyond the 99.5% cutoff and is thus probably an underestimate. As depicted, the average region of facilitation is smaller than the average region of inhibition. - 52 - Regions of Binocular Feedback: Positions Relative to Monocular RF 40 35 30 z 25 C 3 0~ -, S-20 :::io ~ ~ ~ ~ ~ ~ ~ ~ ~ (0 0 S1 15 10 5 Distance from Receptive Field Figure 15 ~ Inhibition ■ Facilitation - 53 - Regions of Binocular Feedback: Size Distribution I Facilitation I ...... 0 .... ~ .0 E :::i z 6 I Inhibition I 1.00 30° Region Size Figure 16 5.0° 6.0° - 54 likely underestimates because they represent only the area over which the binocular difference measured exceeds 99.5% reliability. No difference is found for size distribution or distance to monocular RF for IPSI versus CONTRA, On versus Off, or central versus peripheral. Binocular feedback, it is shown, occurs most often as areas with specific retinal displacement. tation and locations and specific requirements for binocular The majority of cells have multiple regions of faciliinhibition, each with their own location in retinal and depth space. Facilitation tends to occur more often at zero binocular displacement than inhibition. Neither inhibition nor facilitation predominate under the conditions tested. Most regions of BF lie on or near the monocular RF, although a significant number can be more than a few degrees away. Regions of BF can vary considerably in size, possibly from as small as less than O.1° to larger than 5.8°. On the average, inhibitory regions are larger than facilitory. - 55 2.3 Velocity Tuning Thirty-four cells were tested to see if BF varies with of visual sli ts, or edges were used. on Other of the visual stimuli (e.g . , direction of sweep, relative binocular displacement) were sometimes chosen on the basis tests velocity s timuli sweeping over the region around the monocular RF. In general, ver t ical g r atings, parameters the of other the cell, but more often these tests could not be analyzed soon enough, so partly arbitrary values were used. Despite the difficulty in optimiz i ng the visual stimuli for BF, sharp velocity tuning was found to be far more common for BF than for the monocular response of each cell. As with displacement tuning, BDHs were tudes were generated from thirty-two cells. paramet e rs tuned a nd usable A number of cells were made in various of the visu a l s timuli. All BF veloc i ty tunes could be identified a s belon ging to one of the following 1) ampli- Forty-six complete and given more than one ve locity tune with alterations other and extracted from regions of significant BF, in this case at each velocity tested (Figures 17 and 18) 1 . tunes constructed containing five categories: a sharp peak (Figures 20, 21, and 27), 2) tuned and containing a steep trough (Figure 19), 3) untuned and erally upward-going (from gen- strong inhibition t o weak, or from weak Figure 22), facilitation to strong downward-going (Figure 24), or 5) untuned and basically flat (Figure 4) untuned and generally 1. When more than one region of BF was found on a BDH, the maximum amplitude from all regions of one type was used in the velocity tune. When regions of both inhibition and facilitation were found, a "split" tune was made (as in Figures 24 and 27). - 56 23) 2 . The tuning of the monocular responses could be similarly clas- sified, except that no monocular tunes were found that contained troughs. Quantitative criteria were set for establishing the existence of peaks, troughs , upward-only, downward-only, and flat tunes. Visual inspection of a tune was considered inadequate because the appearance of possible peaks and troughs could be exaggerated or diminished by variations in a graph's scale. increase on the lower A peak should have both a maximum velocity side and a maximum decrease on the higher side that exceeds a certain slope. A slope equal to 7% of peak BF per one degree/second was chosen because tunes with multiple small upward and downward fluctuations (as in Figure any the 23) never had of their apparent peaks or troughs meet ing this requirement. As an example, a cell with 100 s/s facilitation at 16°/sec would have to show a facilitation drop-off to less than 44 s/s at 8°/sec (or above) and at 24°/sec (or below) 16°/sec. The criteria to for qualify a as trough having a real peak at is the same, but inverted in direc tion. If a tune had no real peaks or troughs it was classified as upward or downward if the average slope of the tune exceeded 1.25% of the peak BF per deg/sec (such as dropping from 75 at 4°/sec s/s inhibition to at least 100 s/s at 24°/sec). If it did not meet these requirements, it was then classified as flat. Of the forty-six BF velocity tunes, only five (11%) were downward-going, five (11%) were upward-going, and two (4%) were flat. 2. In all figures showing velocity tunes, inhibition is depicted below the X-axis as negative spikes/second and not as a downward slope or curve. - 57 - Figure 17 This figure shows the monocular histograms and BDHs of LGN cell LA5-8 (IPSI-On, RF 5o from AC) as tested by a 1.7° grating moving at velocities ranging from 2°/sec to 14°/sec away from the AC. The filtered BDH is not shown. At 2°/sec, inhibition is very weak, and there is even a small amount of facilitation (slight but in this case significant). Binocular inhibition increases until 12°/sec where it peaks. The tuning is relatively weak in this cell, and it was chosen for depiction because it contains the greatest number of velocities simultaneously tested for any unit. Most often, three to four velocities would be tested at once. An entire tune could consist of from one to four tests of separate sets of velocities. This tune also demonstrates that onset inhibition can occur with gratings and that for this cell the onset BF is not as tuned to velocity as is the response during the sweep. Figure 18 This figure shows the monocular histograms and filtered and unfiltered BDHs for unit LA7-3 (CONTRA-On, RF 5° from AC) as tested by a leading edge moving at velocities 4°/sec to 16°/sec towards the AC. Other velocities were tested and the entire tune can be seen in Figure 20. The "ragged" BDHs are clearer after statistical filtration and the transition from inhibition at 4°/sec to facilitation at 8°/sec back to inhibition at 16°/sec during the sweep becomes clear. Also shown is a strong onset binocular inhibition with some offset inhibition at 4°/sec and possibly some offset facilitation at 16°/sec. Notice that the cell also has a monocular onset and offset response. - 58 - Partial Velocity Tune for Cell LAS-8 Monocular Response to Moving Grating f 1 sec 14 -=~~~~=====,,,====..oc.~~~,:,:::,,r::.,....,..... 12 - - ~ u : . . . . : i . . . L i . ! L W , - : : : : i o : = = = ~ = = = = = = ~ ~ ~ ] ~6-=-~~.L__~~.Lru~=~=~==-====~~- ~ Binocular Difference Histograms I 1 sec Figure 17 - 59 - Partial Velocity Tune for Cell LA7-3 Monocular Response to Moving Slit 300 ms 1 1 o,s 16 4 0 1S J.al:..~~~~~~~.AJI..C------JJ.L-..il.--- Binocular Difference Hist ograms UNFILTERED FILTERED TO p < .005 Figure 18 - 60 - Figure 19 This figure shows the velocity tunes of the monocular response and of the BF of unit LA5-6 (IPSI-Off, RF 4° from AC) to a 1.7° grating moving horizontally over the monocular RF away from the AC. The monocular response is not well tuned to velocity, showing a general increase in response with increased velocity. In contrast, the binocular inhibition is sharply tuned, decreasing in a trough to a maximum inhibition at 12°/sec. In all figures showing velocity tunes, inhibition is depicted below the X-axis as negative spikes/second. Figure 20 Shown are the monocular and BF velocity tunes for cell LA7-3 (CONTRA-On, RF 5° from AC) as tested by a leading edge moving over the monocular RF towards the AC. This monocular velocity tune is categorized as flat. The BF is strongly tuned to velocity, changing from inhibition at low edge speeds to a peak facilitation at 8°/sec back to inhibitition at a higher velocity. (See also Figure 18.) Figure 21 The monocular and BF velocity tunes for cell LA2-5 (IPSI-On, RF 5° from AC) to a slit moving over the RF towards the AC are shown. The monocular tuning consists only of a slow decrease in response with velocity, whereas the BF has a clear peak at 5°/sec. Unlike the unit in Figure 20, the entire range of velocities tested produced binocular facilitation. - 61 - VELOCITY TUNES OF BINOCULAR FEEDBACK AND NORMAL MONOCULAR RESPONSE OF L~~ CELL LA5-6 300 PEAK MONOCULAR RESPONSE S p 200 f k e s I 100 s e C 0 0 40 PEAK BINOCULAR DIFFERENCE 10 20 s ~ -60 k e s I - 80 S e C 100 VELOCITY (Degs/Sec) Figure 19 JO - 62 - VELOCITY TUNES OF BINOCULAR FEEDBACK AND NORMAL MONOC ULAR RESPONSE OF L~~ CELL LA7-3 200 PEAK MONOCULAR RESPONSE s p f k e s I 150 s e C 100 40 s p PEAK BINOCULAR DIFFERENCE 20 1 k e s I 0 e 20 5 10 s C 40 VELOCITY (Oegs/Sec) Figure 20 - 63 - VELOCITY TUNES OF BINOCULAR FEEDBACK ANO NORMAL MONOCULAR RESPONSE OF L,~ CELL LA2-5 150 PEAK MONOCULAR RESPONSE s p 1 k 100 e s I s e 50 C 0 100 PEAK BINOCULAR DIFFERENCE s p I k 80 e s I s e 60 C 40 0 5 10 VELOCITY (Degs/Sec) Figure 21 15 20 - 64 - Figure 22 This figure shows monocular and BF velocity tunes for LGN cell LA6-4 (CONTRA-On, RF more than 15° from AC) to a 1.7° grating moving over the RF away from the AC. This cell is an examp l e of those with BF that show only a slow upward trend as velocit y increases. The same trend is seen with the monocular respons e. Figure 23 Unit LAS-5 (CONTRA-Off, 4° from AC) shows BF that is relatively invariant with velocity and a monocular response that slowly increases with higher sweep speeds. These tunes were done with a 1.7° grating moving over the RF away from the AC. Figure 24 Cell LA2-3 (CONTRA-On, RF 6° from AC) shows BF that shows a slow downward trend with increased velocities and a monocular response with an opposite upward trend. The tests were done wi th a slit moving over the RF away from the AC. The BF measure i s split at 2°/sec because in that sweep, regions of both facilitation and inhibition were found. The values plotted represent the maximum of each seen in the sweep. This "splitting" was not unconnnon. - 65 - VELOCITY TUNES OF BINOCULAR FEEDBACK ANO NORMAL MONOCULAR RESPONSE OF LG~ CELL LA6-4 PEAK MONOCULAR RESPONSE S 200 p f k e 5 I 150 s e C 100 5 PEAK BINOCULAR DIFFERENCE S 40 p 1 k e s / s 15 10 - 60 e C 80 VELOCITY (Oegs/Sec) Figure 22 20 25 - 66 - VELOCITY TUNES OF BINOCULAR FEEDBACK ANO NORMAL MONOCULAR RESPONSE OF LG~ CELL LA5-5 200r- PEAK MONOCULAR RESPONSE S p 1 k 150 e ' s 100 e C 50 01.r-----~-----~-----~----~~----~-------' 10 20 30 PEAK BINOCULAR DIFFERENCE s p 1 k -50 e s I s e 100 C 150 VELOCITY (Degs/Sec) Figure 23 - 67 - VELOCITY TUNES OF BINOCULAR FEEDBACK ANO NORMAL MONOCULAR RESPONSE OF~~~ CELL LA2-3 200 PEAK MONOCULAR RESPONSE 5 p 1 k 150 e s I s e 100 C 50 50 5 p 1 k 0 PEAK BINOCULAR DIFFERENCE 10 20 e s I s e 50 C 100 VELOCITY (Oegs/Sec) Figure 24 30 40 - 68 - The majority, 33 (72%), had real peaks and/or troughs. Of these, thirteen (40%) had a single peak, fourteen (42%) had a single trough, five (15%) had both a peak and a trough (as in Figure 25), and one (3%) had a trough, a peak, and then a trough 26). Of the (Figure twenty-one trough minima identified, nineteen were inhibitory and two were at zero (and thus were troughs Seventeen of the in a range of facilitation). nineteen peak maxima were facilitory and two were inhibitory. It should be noted that in many cases the velocity for spanned both facilitation and inhibition (as in Figures a cell 20, 24, 25, 26 and 27). tune In almost every case, the BF for each indi- vidual cell was strikingly more precisely tuned for velocity than its monocular response to the same stimuli (as in Figures 19, 20, 21, 25, 26 and 27). BF tends to be best tuned at a particular range (Figure 28). Peaks and troughs occur far more often 12°/sec any than other 8°/sec. No difference in interval, with distribution is of velocities between 6 and 19 (48% of 40) falling at seen between peaks and troughs. An average BF amplitude was calculated for each of the BF tunes containing no peaks or troughs (and untuned). Twelve (92%) of these were inhibitory (as 23, and 24). All of the BF tunes downward-going, forms most often seen in primarily to 71 s/s. that were monocular thus in thirteen considered Figures 22, upward-going or responses, were inhibitory, with the averages for each tune ranging from 7 - 69 - Figure 25 The monocular and BF velocity tunes for LGN cell LA6-3 (CONTRA-On, RF more than 15° from AC) taken with a 1.7° grating moving across the RF away from the AC. The monocular tune is upward-going whereas the BF tune has both a peak (at 14°/sec) and a trough (at 24°/sec). The point at 14°/sec is just at threshold (see Text) for classification as a real peak. Figure 26 This figure shows the monocular and BF velocity tunes for unit LA5-7 (CONTRA-Off, RF 8° from AC) for a 1.7° grating moving over the RF away from the AC. This cell was unique in that it shows a trough, a peak, and then a trough. Notice also that the tune includes both facilitation and inhibition. The monocular tune is of the typical upward-going type. Figure 27 Shown here are the BF and monocular velocity tunes for cell LA7-1A (CONTRA-Off, RF 4° from AC) for a 1.7° grating moving across the RF away from the AC. The BF is split at 6°/sec because both facilitation and inhibition were found at that velocity. The monocular tune is one of three that has a real peak, which is nevertheless fairly weak. - 70 - VELOCITY TUNES OF BINOC UL AR FEEDBACK ANO NORMAL MONOCULAR RESPONSE OF L~~ CE LL LA6-3 300 PEA K MONOCULAR RESPONSE s p 200 1 k e s I s 100 e C 0 PEAK BJNOCUL AR DIFFERENCE 50 s p 1 k e s I s e C - 50 VELOCITY (Degs/Sec) Figure 25 - 71 - VELOCITY TUNES OF BINOCULAR FEEDBACK AND NORMAL MONOCULAR RESPONSE OF LG~ CELL LAS-7 80 60 PEAK MONOCULAR RESPONSE s D 1 k 40 e s I s 20 e C 0 -20L 20 s D 1 k e s PEA K BlNOCULAR DIFFERENCE 0 - 20 I s e C - 40 60 VELOCITY (Degs/Sec) Figure 26 - 72 - VELOCITY TUNES OF BINOCULAR FEEDBACK AND NORMAL MONOCULAR RESPONSE OF L~~ CELL LA7-1A 80 PEAK MONOCULAR RESPONSE s D 1 k e s 60 I s e 40 C 40 s p PEAK BINOCULAR DIFFERENCE 20 1 k e s I s e C 20 VELOCITY (Oegs/Sec) Figure 27 - 73 - Figure 28 This fig ure shows two ways of representing the velocities at which peaks and troughs were found (called here "critical" velocit ies). In both cases, a c l ear concentration of critical velocities is within the 6°/sec to 12°/sec range with a smaller secondary grouping around 24°/sec. Also, in both cases no significant difference is seen between peaks and troughs. Figure 28-A: Simple histogram showing the number of peaks and troughs that were found for each velocity bin (which are 2°/sec wide). Figure 28-B: Velocity tunes were done at often differing test points. A peak at 8°/sec might be isolated between points as close as 7 and 10°/sec or as far apart as 4°/sec and 16°/sec. Each peak and trough could be better identified by a confidence interval of velocities over which it falls (such as 7.5°/sec to 9°/sec versus 6°/sec to 12°/sec). This histogram shows the number of peak and trough confidence intervals that contain each velocity. The broadening of this histogram over the one above results from including the range of uncertainty for each point. Even with this correction, the concentration of critical velocities is clear. - 74 - VELOCITY TUNING OF BINOCULAR FEEDBACK: Distribution of "Critical" Velocities ■ PEAKS ~ TROUGHS 20 15 10 V'l 5 ~ O' ::::; 0 ... 1"'0 4-5 c::: 12-13 8-9 16-17 20-21 (';l V'l ~ 25 (lJ - 0.. 0 ~ 20 B (lJ ..0 E ::::; z 15 10 5 5 10 15 20 Velocity ( degs /sec) Figure 28 24-25 28-29 - 75 It was possible in one case to do a BF different velocities (Figure 30). location map of the various regions three In general, the overall organiza- tion 9f the map was similar at different velocities, but sity at the inten- of BF changed considerably. It also appears that the preferred displacement for a region of BF could be sh~fted at different velocities, although many more examples would be needed to clearly demonstrate this. No correlation of the . occurrence troughs, of tune type, IPSI versus CONTRA, On or amplitude of peaks and or of tuning steepness could be found with versus Off, or central versus peripheral cells. In summary, BF was found to be considerably velocity changes (Figure 29). than are more sensitive to monocular responses to the same stimuli Further, BF is most sensitive to changes occurring in the range of 6°/sec to 12°/sec (with possibly a secondary grouping at 24°/sec). Cases of BF that are similar to the common monocular file being poorly tuned to velocity are shown in general to con- of sist mostly of inhibition. pro- - 76 - Figure 29 This figure shows all of the velocity tunes taken. The upper figure contains all of the monocular tunes and the lower shows all of the BF tunes. The scaling is the same for both. A strong difference can be seen in that the monocular tunes are for the most part linear, giving a fairly smooth and even appearance. The lower figure consists of many tunes with slopes that change rapidly with velocity, particularly around the 6 to 12°/sec range, thus giving a more ragged appearance. A comparison of the two yields a striking demonstration suggesting a difference in the processing of velocity information between a monocularly stimulated LGN and a binocularly stimulated LGN. - 77 - ALL VELOCITY TUNES OF MONOCULAR RESPONSES COMBINED ----------·- •oo 300 s D 1 k e s I s 200 e C --·-- ------ ------__--:.:--··.: ' ,:•_ _ .. ... .. -· --------- ;;L;:~- - ----------- 100 ·- ·- ·-·- ·- ·- ·- (" . ,, --- -- ✓✓~-~~~~~~~~~~ ·····~·-···~ · ·•·· ~= - --=·-----~------- = ··-= ---· · 00L.__::.::.::...._..___ _ _1_._o_ _ ___._ _ _~20=------'------::1'=-o----'-----;;,o VELOCITY (degs / secl ALL VELOCIT Y TUNES OF BINOCULAR DIFFERENCES - COMBJNED too 100 s D - -:- I k e s I 0 40 s e C 100 I ·' / VELOCITY ( degs/secl Figure 29 - 78 - Figure 30 This figure shows three BF location maps taken at different velocities for LGN cell LA9-3 (IPSI-Off, RF 13° from AC) taken with a slit moving across the RF towards the AC. At least two islands of facilitation a r e seen in each map, with both showing slight shifts in optimal position in visual space. The amplitudes of these shifts are not large and are possibly due to eye movements or other measurement errors. The shift in amplitude of the right island of facilitation to a maximum at 8°/sec cannot be accounted for by such errors. The inhibitory regions also show similar changes with velocity, with a maximal area of distribution at 4°/sec. The onset and offset BF are included and also change considerably with velocity. The depiction of onset and offset BF is as described in Figure 12. - 79 - MAP OF BINOCULAR FEEDBACK: EFFECT OF VELOCITY FOR CELL LA9-3 FACILITATION D P <.005 [ p ;:;:.005 ON E 2· "° .E 0 0 8 TO 17 SI. • 18-34 S/9 • 35-50519 ~ 0,a I c:: 0 z c:: ~ 0 . . , c:: ... -2· 0 </I 0 0.. <U > 0 4 ~ ...... a. o,a ,c!:!.,,.. ,I..=. - ...... 1· ;I: <II <> -1· . .. .... ... ,~ ----~ .... o· -1° Relative Binocular Displacement Figure 30 1· C_'-j 11 ON ...c:: INHIBITION 8 :_-_-g - 80 3.4 Directional Selectivity Nineteen cells were t ested to see if BF v ar ie s wi th the direction visual stimul i swe eping over t he monocular RF. Vertical gratings, slits, and edges moving horizontally were used to for of compare BF found sweeps moving t owar d the AC (notated as "180° 11 ) with sweeps mov- ing away from the AC (0°). varied parameters Often directional selectivity tests with of the visual stimuli were done on the same cell. Twenty-seven such direction tunes were made in all. Four cells were studied to see how the BF location maps vary and seven were tested to see how velocity t uning varies with sweep direction. BF in the LGN was foun d to be very sensitive to sweep direction. BF could vary with sweep direction in one of four ways. shows the distribution of these four types. The most common type ("Change in Amp l itude") refers to either an increase or the Table 2 decrease in magnitude of the BF with a change i n sweep dir ection, as seen in eight cells ( 42% of 19) and in eleven tunes (41% of 27). The second most common showed the total absence of BF a t one direction and clear BF for the oth i~r ("BF - Only One Dir"). This was found in five (2 6% of 19) and in eight tunes (30% of 27). The third involved cells that had multiple regions of BF with at least one region present one cells direction and not for the other ("Extra Region"). (16 % of 19) and five tunes (19% fourth and most striking kind of 27) showed this for Three cells pattern. The of directional selectivity was the reversal of the polarity of BF, such as from inhibition to facilita- tion (Figure 31). Three cells (16% of 19) and three tunes (11% of 27) - 81 - Figure 31 This figure shows the effect of sweep direction on the BF of LGN cell LA4-3 (CONTRA-On, RF 1.50 from AC). A 1.7° grating was swept at 8°/sec over the RF toward the AC (180°) and away from the AC (0°). The BDHs show binocular inhibition during the sweep for 180° switching to facilitation at o0 • The monocular sweep response does not change significantly with direction. Also note the presence of a strong monocular and BF onset. The onset is mostly inhibition, and does not switch polarity. - 82 - • Binocular Feedback as a Function of Sweep Direction for Cell LA4-3 Monocular Response 180° Binocular Difference Histograms FILTERED TO p < .01 UNFILTERED Figure 31 - 83 showed this type ("Polarity Reversal"). Change in Amplitude: !BF - OnlylExtra !Polarity 0-20%l21-40%l41-60%l61-80%l>80%I One Dir IRegion!Reversal -------------------------------1---------l------l-------o 2 2 3 1 I 5 I 3 I 3 = 19 cells 1 4 2 3 I 1 I 8 5 I 3 = 27 tunes Occurrence and Magnitude of Directional Selectivity Table 2 The occurrence and magnitude of directional selectivity of not be correlated in BF could any way with IPSI versus CONTRA or On versus Off. It was also found that overall, neither inhibition nor facilitation was more prevalent for sweeps toward or away from the AC. Further, an increase in BF amplitude was not seen to be more common for a reversal in either direction. o0 For four cells, full BF location maps were made for both 180° sweeps (Figure 32). in All four cells showed significant the organization of these maps. was changed the occurrence and and shifting In general, when sweep direction magnitude of regions and/or the predominance of a particular polarity of BF would also change. Seven cells were given full velocity tunes for and six showed clear differences (Figure 33). cases, reversal both In each of directions these six of direction resulted in shifts of the types listed in Table 2, but not uniformly at all velocities. For another BF region appeared, but only at some velocities. showed large increases in BF amplitude at some two cells, Three cells velocities with - 84 - Figure 32 This figure shows complete BF location maps taken at two different sweep directions for LGN cell LA9-10 (IPSI-Off, RF 11° from AC) using a slit travelling over the RF at 8°/sec. The entire arrangement of BF is shifted by changing sweep direction. The most striking changes are the increased prevalence of facilitation and the loss of a region of inhibition with the shift from 180° to o0 • This type of selectivity is referred to as "Extra Region" in Table 2. The binocular displacement was controlled by a variable prism placed over the dominant eye. The dotted line shows the position of the peak monocular response (and thus the RF). - 85 - LOCATIONS OF BINOCULAR FEEDBACK: Effect of Sweep Direction for Cell LA9-10 FACILITATION INHIBITION □ p< .005 { P~ .005 ~tst1 TO 16 SIS - 17 - 31 SIS f51 - 32-47 s.,s ~1§ C: 0 50 C: 0 ..... r..r, ~ J?. ·~11. II.I II.I 3: "" ..... 10 V') -1 0 -30 0 -5 -40 Relative Binocular Displacement Figure 32 c:J r:_-J - 86 - Figure 33 This fig u re shows the monocular and BF velocity tunes for cell LA2-5 (IPSI-On, RF s0 from AC) to a slit moving over the RF for sweeps toward (180°) and away from (0°) the AC. The monocular tuning consists only of a slow decrease in response with velocity for both directions. The BF has a clear peak at 5°/sec for 180° and a clear trough at s0 /sec for 0° (as well as a region of low facilitation). In this case, the critical velocity was the same for both sweep directions even though the polarity and amplitude of the BF varied. - 87 - 200 VELOCITY TUNES OF BINOCULAR FEEDBACK AND NORMAL MONOCULAR RESPONSE OF L~~ CELL LA2-5 PLOTTED AT OPPOSITE SWEEP DIRECTIONS PEAK MONOCULAR RESPONSES o - s □ p 1 k e s 100 8--.-g 8 o-----R□1...._-------=-------~-e-e--------!I ,O:r--------- I -£l e 0 PEAK BINOCULAR DIFFERENCES 100 s ~ (A. C. -->) 8 s C o• - 1so° CA.C.<--l 50 k e s I s e C 50 VELOCITY (Degs/Sec) Figure 33 - 88 direction change and one cell showed a switch in polarity only at one velocity (Figure 33). Orientation tunes were attempted for eight cells. However, as described in Section 3.2, BF often occurs in multiple regions distributed over the retina around the area of the monocular stimuli swept so Oriented as to totally cover the monocular RF at all sweep angles do not necessarily cover all regions Although RF. of BF for that cell. the bi-directional sweeps described above stimulate exactly the same region of the retinas with a stimulus of the same angle (but a different polarity of direction), this is not true for an orienta- tion tune. for a tune It would be possible to mask off all visual space except particular region of BF and then to do a complete orientation over required that region. However, the computation and analysis to discover from experimental runs the size and location of each region was impractical in the time a cell was normally held. One would expect simple orientation tunes without such corrections to produce multi-modal results with amplitudes and even existences of BF reg ions that vary non-uniformly with angle of sweep. This is what was found for these cells, and orientation tunes were abandoned for study. this However, BF did vary in every case with orientation, if non- uniformly. Figure 4 shows the BF of one cell at four different of directional orientations. In summary, all cells tested selectivity for BF, the polarity of the BF. showed a degree ranging from changes in amplitude to shifts in Further, directional selectivity varied at - 89 different velocities and at different locations on the retina and binocular displacement space. in - 90 3.5 Other Observations Concerning Binocular Feedback One striking feature of many BDHs was the significant BF at the existence of strong and onset and/or the offset of the sweep of the visual stimuli, as mentioned above. From forty-two cells, forty-one showed onset and/or offset BF under some conditions (for example, see BDHs in Figures 10, 17, 18, and 31). The source of these phenomena is unclear and for gratings could simply result from the flashing on of the stimuli over regions of BF. However, in the twenty-seven cells tested with slits, whose sweeps start and stop up to 8° away from the RF, eighteen (67%) cells showed onset offset BF, and BF, twenty-two (81%) fifteen (56%) showed both. Many cells also showed a monocular onset and/or offset response (also in Figures 10, and 31). It and response. 17, 18, was found possible for either to occur with or without the other. Four (15% of 27) showed onset BF with no firing, showed six Four (22%) cells showed (15%) onset monocular offset BF with no similar monocular showed monocular onset with no BF, and two (7%) showed monocular offset with no BF. Fourteen (52%) had both onset responses and sixteen (59%) had both offset responses. It is likely that these responses do not result only from tered Often light. the onset and/or offset BF would occur only at particular binocular displacements (Figures 11, 12, 13, and at particular the 30) and velocities. In some cases, they would have a velocity tuning different from each other and from that of during scat- sweep (Figure 30). the BF occurring As with RF located responses, the onset/offset monocular responses tended not to be well tuned to velo- - 91 city. This type of BF is similar in amplitude to the above described sweep BF, with the average maximum (for each cell) onset facilitation at 66 s/s, onset inhibition at 63 s/s, offset facilitation at 57 s/s, and offset inhibition at 66 s/s (compare to Figure 6). Unlike sweep BF, which shows approximately equal inhibition and facilitation, onset 9 of = 27 33%). (15 of (10 = 27 57% More than three times as many cells show predominantly onset inhibition than show predominantly onset tation of binocular inhibition occurs in almost twice as many cells as onset facilitation versus occurrence facili- of 27 = 37% versus 3 of 27 = 11%). No such asymmetry was found for offset BF. Differences between IPSI and CONTRA or On and mentioned. With Off for this type of BF could not be substantiated. Another type of observation about BF should be the set-up used, a crude, wide bin histogram showing monocular and binocular responses in tandem was produced in experiment. It was often the strong differences between the would diminish as more author's monocular trials were real time during an impression that initially and binocular histograms run. The possibility that BF fatigues was looked at more closely for two cells. A velocity tune was run for unit LA4-4 for 28 minutes, and the BDHs were compared for the first and last 14 minutes. This unit showed a region of facilitation at 4, 8, and 16°/sec and a region of inhibition at 4 and 8°/sec during the first 14 minute run. The second 14 minute run showed only the inhibitory region, and only at 4°/sec. A similar test was done on unit LA6-4, with a direction tune run continuously for almost a - 92 half-hour. Initially, two regions of facilitation and one of inhibi- tion was seen. disappeared. By 11.3 minutes, the regions of facilitation had The tune was continued and the facilitation did not re- appear within the remaining 13 minutes. Further, the initial ampli- tude of inhibition seen was reduced on an average of 23% to the final values seen in the last 13 minutes. fatigue, that This is suggestive that BF can inhibition may be more resistant to fatigue, and that long runs testing for BF and averaging over the whole run may produce underestimates of its magnitude. It should be noted that in about one half of at, the cells looked the response of the non-dominant eye alone to the visual stimuli was also taken together with the dominant and In general, of the amplitude of modulation binocular responses. spontaneous activity by stimulation of the non-dominant eye alone was from O s/s to 8 s/s and could not account for the BF found. Finally, one cell produced a BDH with (Figure 34). a very unusual pattern Unfortunately, the data about the cell and the run were scrambled by the computer and lost. not statistically significant. Further, the unusual effect was It is included only for completeness. In summary, BF occurring at the onset and offset of a slit sweep BF. Monocular onset/offset responses were also connnonly seen, and either monocular was or common binocular other. and similar in amplitude to sweep responses could occur alone or in combination with the These effects could also be located in binocular displacement space and have particular velocity requirements. Some evidence exists - 93 - Figure 34 This figure shows the monocular responses and BDHs for a cell that was probably an On unit, tested with a slit. The cell and run data were scrambled by the computer and lost. The unusual effect seen in the unfiltered BDH was not statistically significant. The figure is included only for completeness. - 94 - Binocular Feedback Pattern for Cell LA6-? Monocular Response Binocular Difference Histograms 3oom. 1 FILTERED TO p > 00.5 UNFILTERED ....._.......,. __ Figure 34 I - 95 that for onsets, inhibition Further, some evidence was is more prevalent generated that than facilitation. suggests that BF can fatigue over the order of minutes, and that facilitation may be sensitive than inhibition. Lastly, pure non-dominant visual stimula- tion was excluded as a significant source BDHs. more of the effects seen in - 96 5. DISCUSSION The vast majority of LGN relay cells showed both binocular tion facilita- and inhibition when tested with oriented stimuli that were sys- tematically altered along parameters that profoundly affect unit firing. The amplitudes of these interactions were strong; in as many as half of the cells studied they exceeded normal firing 50% of the prevalent than cell's The typical LGN cell had multiple regions of rates. binocular feedback (BF) with facilitation only slightly and cortical inhibition. less strong Regions tended to be very close to the cell's monocular RF on the retina, with a maximum distance of away. Regions ranged in 6° size from 0.1° to 6.0°, with facilitation averaging 1.4° and inhibition 2.0°. The BF found in this parameters of visual way was highly tuned to a number of stimuli. The majority of regions had specific requirements for retinal disparity. Facilitation was optimal at zero disparity twice as often as was inhibition, and both had more regions requiring convergent disparities than divergent. In most BF was occurred in the interval between and 12°/sec. Monocular responses were not well tuned to velo- city. Further, most amplitude, BF was disappearance directionally selective. orientation, but Changes in of regions of BF, and even polarity rever- sals were seen with a 180° change in sweep direction. stimulus the well tuned to velocity, with peaks and troughs seen in equal numbers. Most peaks and troughs 6°/sec cases, BF varied with in a non-uniform fashion. Technical con- siderations made conclusions about orientation tuning difficult (see - 97 RESULTS). Although BF was found to visual stimuli, no be significant sensitive differences between CONTRA and IPSI cells and between indication was found that On to many were and Off that fatigue the BF can cells. Some facilitation is stronger in the retinal Evidence was presented of of seen in any way periphery than in the cente r. some variables in that suggests the order of minutes. phenomenon of sweep onset and offset BF was found and described. The - 98 5.1 Relationship to Previous Work Various new responses techniques were used in study. primarily on responsivity. non-dominant eye alone Earlier studies the response of (Singer, 1970; Sanderson et al., 1971; Fukuda and Stone, 1976; Rodieck and Dreher, 1979). were binocular a sequence of a monocular run, then binocular, then monocular, and so on to find BF, or simply on the LGN were taken in tandem with monocular responses to eliminate errors due to variability in neuron relied this Here, histograms constructed of the full sweep of the difference between a bino- cular and monocular response of a cell, thus showing BF which doesn't necessarily fall during the peak firing of the cell. Statistical filtering was used to identify intervals of significant BF. studies were conducted with oriented stimuli Finally, controlled for parameters that influence cortical activity. In general, the present study found BF more often facilitation) and (particularly at higher amplitudes. Sanderson et al. (1971) for example, shows inhibition from the non-dominant eye on the order of 5 s/s to 10 s/s, whereas this study found average inhibition and facilitation amplitudes of SO s/s to 70 s/s. In light of the finding BF is very tuned to various aspects of a visual stimulus and that this study involved systematic searches through these aspects, it not suprising that is more BF and larger amplitudes were found. Other problems could have prevented earlier workers from seeing as much as that BF is present. Typically, in past studies the maximum firing rate of the binocular response was compared to that of the monocular response - 99 - so differences that did not fall at the time of peak firing could have been missed. In addition, evidence was presented here that BF fatigues. Most of our runs were done with just enough trials to get good statistics. This may not works, and some fatigue could have been true in the earlier have diluted measured amplitudes. Thus, there are several reasons why earlier studies did not see as much BF as is reported here. Suzuki and Kato (1966) and Suzuki and Takahashi (1970) binocular inhibition occurring for IPSI cells three to four times as often as for CONTRA whereas the present between IPSI work found no distinction or CONTRA in any way related to BF. They used electric shocks to optic nerves to workers reported test for binocular interactions. found a much smaller asymmetry, if any, using visual stimuli (Sanderson et al., 1971; Singer, 1970; and Rodieck and Dreher, and, Other as 1979) stated above, earlier reported BF was of low amplitude. The current work used exclusively visual stimuli and found high amplitude BF. It shows is no easily conceivable that optimal visual triggering of BF asymmetries while suboptimal triggering or non-visual stimuli could emphasize a small IPSI over CONTRA bias . The des cr iption of arrangements of fields of BF extends works. Sanderson's inhibition fell primarily earlier on the same retinal position as the monocular RF; it is shown here that this is also true for facilitation, and that both types of BF occasionally lie as far as 6° away from the RF. Because recorded in this study, both positions are simultaneously it is difficult to ascribe this result to - 100 errors in measuring eye position. Sanderson et al. found fields from 1.5° & Singer (1977) found facilitory to 6° wide and regions 0.5° to 2° wide. Schmielau The current report is similar. However, the size of inhibitory regions reported by Schmielau and Singer was 5° to 10°. Inhibitory regions found here with oriented stimuli averaged 2° wide and never exceeded 6° and multiple regions for each cell were commonly found. Their study was surprising they that done with flashed spots. It is were able to find cortico-geniculate feedback using stimuli that typically do not drive cortical cells. The current results were obtained with stimuli optimized for cortical neurons and it is not unlikely that they would reveal a different arrangement BF. of It is conceivable that their technique grouped multiple regions into one. The present study, in general, extends the concept of BF to include multiple regions that are well tuned to visual parameters not usually thought to be important in the LGN. - 101 5.2 Source of Binocular Feedback The earlier study of Schmielau and Singer most of (1977) demonstrated that the binocular facilitation and some of the inhibition found in the LGN under their test conditions disappeared cortices were cooled. when the visual This work also suggests that BF is cortical in origin. Tight tuning of cortical units to relative retinal disparity has been demonstrated in the cat (Barlow et al., 1967; Pettigrew et al., 1968b; Bishop et al., 1971), in the monkey (Hubel and and in sheep Wiesel, 1970) (Ramachandran et al., 1977). This work shows that the majority of BF regions also have tight tuning to retinal disparity, which suggests input from the visual cortex. Neurons of the visual cortex have been shown to be tuned to velocity of visual stimuli (Pettigrew, 1968a). Recent studies gen- erally agree that simple cells are tuned to reports the lower velocities, with of 2°/sec to 4°/sec as the average optimal, and that complex cells are tuned to higher velocities, 16°/sec to 18°/sec as the average optimal velocity (Movshon, 1975; Gilbert, 1977; Goodwin and Henry, 1978; Hess, 1979). The greatest number of cortical both were categories combined tuned 12°/sec. Studies of cortical 1977; BF, 6°/sec layer VI cells indicate that the cortico-geniculate system includes both (Gilbert, of to velocities from 5°/sec to 10°/sec~ which is very close to the range found here for to cells simple and complex units. Harvey, 1978). It is likely that BF that comes from the cortico-geniculate system would reflect the tuning properties of - 102 - layer VI cells, which have been specifically confirmed to be well tuned to velocity (Leventhal and study and earlier Hirsch, 1978). Both the present studies confirm that for the most part LGN cells show poor or no velocity tuning when tested monocularly and it is thus unlikely that interlaminar input is responsible for the BF found (Dreher and Sanderson , 1973; Hess and Wolters, 1979; and in the mon- key, Lee et al. t 1979). The comparison of untuned monocular responses with well tuned BF (see RESULTS) is striking. Further, the report Daniels et al. (1977) that by small orientation biases in LGN cell response disappear at stimuli velocities above 20°/sec is consistent with this interpretation. Many cortical cells have also been shown to be ally selective very direction- (Pettigrew et al., 1968a) and Gilbert (1977) reports that layer VI cells are the most consistently directionally selective cells in area 17. The current finding that BF is also directionally selective lends further support for the notion that it is cortical in origin. The evidence that BF cortically mediated multi-synaptic amplitudes of the can fatigue is also consistent with feedback. Finally, the BF found with cortical trigger features with the current techniques far exceed the amplitudes of interlaminar tion found earlier to inhibi- (that survives cortical cooling - Singer, 1970; Sanderson, et al., 1971). able maximum The simplest explanation currently avail- account for the data is that the BF found originates in the visual cortex. - 103 It is of course posssible that this BF doesn't originate in cortex, the but instead in some other area. The cortico-geniculate input is an anatomically dominating feature of LGN structure (see below) and i t i s t he most plausible source. Future studies would best demonstrate that co rtical cooling el iminat e s the kinds of BF described. It should be noted that some BF was found that was interlamina r in origin and not most likely cortical. A small number of cells showed BF that was not tuned to velocity and in twe lve out of teen Interlam- cases it was inhibitory (see Figures 22, 23 and 24). thir- inar inhibition is well substan tiated and one would e xpect it not be more veloc i ty tuned than LGN monocular responses. Binocular dis- placement tunes also revealed a few cases of BF that could easily interlaminar. As described in RESULTS, regions to be of inhibition or facilitat i on falling solely on the non-dominant retina would show up as a band on the BF location maps. Five such bands were seen and four of them were inhibitory. That the current techniques suggest inter- laminar inhibition makes it unlikely that some kind of technical or systematic error is responsible for data suggesting a cortical origin for the other BF. The data suggest a few ideas about the organization of from the visual cortex to the LGN. First, it is possible that multi- ple cortical units functionally converge on each LGN cell. units feedback showed cells LGN multiple regions of BF, each with different polarities and retinal disparity requirements. To cortical Most that are date, no one has reported tightly tuned to more than one disparity. - 104 Also, some cells showed BF that changes polarity with a change in sweep direction. It seems less likely that one cortical unit could be both facilitory and inhibitory than that more converge on than one unit could each LGN cell. Second, there is no evidence that the BF found is related to the CONTRA-IPSI or On-Off classification, as if the incoming cortical fibers make no distinction between these types. From both of these ideas, an image emerges of cortical fibers con- verging on LGN cells with little specificity, other than topographic. - 105 5.3 Functional Significance The cortico-geniculate feedback system is very large, with more than half the cells in layer VI of area 17 projecting to the LGN and about half of the synapses in the LGN of the type that degenerate when cortex is removed and Kelly, 1975). the (Jones and Powell, 1969; Guillery, 1971; Gilbert Only 30% of LGN synapses are retinal in origin. It is likely this system plays a very important role in visual information processing; however, this role is still unclear. Schmielau and Singer (1977) have suggested that the system may be used to highlight a distinction between foreground and background. This ported by idea is sup- the current evidence that facilitation is found much more often than inhibition at zero retinal disparity. This would have effect of increasing the LGN's response to features on the plane of fixation while inhibiting responses to features not Even the though on that plane. such corrections could be done in a higher center where individual neurons are more directly responsive to both eyes, there could be an advantage to feeding back binocularly derived information to a relay center that has images. The idea not yet fully mixed been two retinal that the LGN acts as a general preprocessing area upon which modifying feedback arrives from many already the reviewed by Singer higher centers has in 1977 and by Burke and Cole in 1978. The finding that BF shows increased facilitation inhibition and increased at velocities 6°/sec to 12°/sec might be explained in the following way. Visual features moving at these velocities in partic- - 106 ular would be more strongly facilitated when on the plane of fixation and more strongly inhibited when off of that plane than features moving at other velocities. Pritchard and Heron (1960) reported that a cat's eyes show "flicks" averaging (1959) showed that stabilized 13°/sec, images and Ditchburn et al. on human retinas preventing similar flicks (4-0 /sec to 8°/sec) destroy image i.ntegrity. Although in the Pritchard and Heron preparation these flicks were rare, freely moving cats might show more frequent flicks. be The BF mechanism could designed to maximize highlighting of the plane of fixation during these eye movements. In general, the eyes are constantly moving and deal with creating changing images. velocity and a single brain direction feature and could be involved in other ways in this interested in the velocities the to measure the depth of a feature (Pettigrew, et al., 1968a). Finally, the velocity and direction specificity of BF might be portant of in addition to or instead of eye movements. It might be part of the difficult task of stimulus matching that is necessary for cortex must world image from two rapidly The BF mechanism is clearly concerned with processing. It could be directly objects stable the to the unim- functioning of the cortico-geniculate system and be simply an artifact of the specificity of the cortical units involved. It is felt that the function of the cortico-geniculate system is still unclear. It is possible that the description of characteristics of the system given here is essentially complete, but it is not clear to the author why such a massive system would evolve for the limited, - 107 though important roles suggested above. Perhaps of the nature elucidation and functioning of the areas that receive input from the LGN will illuminate the importance here. further of the mechanisms detailed - 108 References Barlow, H.B., Blakemore, C. & Pettigrew, mechanism J.D. (1967). of binocular depth discrimination. The neural J. Physiol. 193: 327-342. Bishop, P.O. & Davis, P. (1953). Bilateral interaction in the lateral geniculate body. Science. 118: 241-243. & Vankur, Bishop, P.O., Kozak, W., Levick, W.R. determination of the projection G.J. (1962). The of the visual field onto the lateral geniculate nucleus in the cat. J. Physiol. 163: 503- 539. Bishop, P.O., Henry, G.H. & Smith, C.J. (1971). Binocular interaction fields of single units in the cat striate cortex. J. Physiol. 216: 39-68. Burke, W. & Cole, A.M. (1978). Extraretinal influences on the lateral geniculate nucleus. Cleland, B.G., recording Dubin, on Rev. Phys. Bioch. Phann. & Levick, M.w. input and W.R. 80: 105-166. (1971). Simultaneous output of lateral geniculate neurons. Nature 231: 191-192. Coenen, A.L. & Vendri k, A.H. (1972). Determination ratio of cat's of the transfer geniculate neurons through quasi-intracellular recordings and the relation with level of alertness. Res. 14: 227-242. Daniels, J.D., Norman, J.L. & Pettigrew, oriented moving bars J.D. (1977). Biases for in lateral geniculate nucleus neurons of normal and stripe-reared cats. Exp. Br. Res. 29: 155-172. - 109 - s. (1959). Ditchburn, R.W., Fender, D.H. & Mayne, trolled movements of the retinal image. Vision with J. Physiol. con- 145: 98- 107. & Sanderson, Dreher, B. rsponses K.J. (1973). Receptive field analysis: to moving visual contours by single lateral geniculate neurons in the cat. l.• Physiol. 234: 95-118. Dubin, M.W. & Cleland, B.G. (1977). Organization of visual inputs interneurons rophys. of lateral geniculate nucleus of the cat. to J. Neu- 40: 410-427. Duckworth, W.E. (1968). Statistical Techniques in Technological Research, PP• 39-43. London: Methuen & Co. Fukuda, Y. & Stone, J. (1976). Evidence influences on X- geniculate nucleus. and of differential inhibitory Y- type relay cells in the cat's lateral -Br. -Res. - 113: 188-196. Gilbert, C.D. (1977). Laminar differences in receptive field ties of cells in cat primary visual cortex. proper- J. Physiol. 268: Gilbert, C.D. & Kelly, J.P. (1975). The projections of cells in dif- 391-421. ferent layers of the cat's visual cortex. l.• Comp. Neurol. 163: 81-106. Goodwin, A.W. & Henry, G.H. (1978). The influence of city stimulus velo- on the responses of single neurones in the striate cortex. l.• Physiol. Guillery, R.W. lateral 277: 467-482. (1967). geniculate visual cortex. Pattern of fiber degeneration in dorsal nucleus of the cat following lesions in the l.• Comp. Neurol. 130: 197-222. - 110 Guillery, R.W. (1971). Patterns of synaptic interconnections dorsal lateral geniculate review. Vis. Res. Sup. nucleus the of cat and monkey: a brief 3: 211-227. Harvey, A. R. (1978). Characteristics of corticothalamic area 17 of the cat. in Neurosci. L■ neurons in 7: 177-181. Hendrickson, A.E., Wilson, J.R. & Ogren, M.P. (1978). The neuroana- tomical organ ization of pathways between the dorsal lateral geniculate nucleus and the visual cortex of old world and new world primates. l.• Comp. Neurol. 182: 123-136. Hess, R. & Wolters, W. (1979). Responses of single cells in cat's lateral geniculate nucleus and area 17 to the velocity of moving visual stimuli. Exp. Br. Res. 34: 273-286. Hollander, H. (1972). Autoradiographic evidence for a projection from the striate cortex to the dorsal part of the lateral geniculate nucleus in the cat. Res . -Br. -- 41: 464-466. Hubel, D.H . & Wiesel, T.N. (1961). Integrative action lateral geniculate body. Hubel, D.H. & Wiesel, interaction, cortex. and T.N. l.• Physiol. (1962). functional l.• Physiol. the cat's 155: 385-398. Receptive architecture fields, binocular in the cat's visual 160: 106--154. Hubel, D.H. & Wiesel, T.N. ( 19 70) . monkey. in Ste reoscopic vision in Macaque macaque lateral Natur e 225: 41-44. Hull, T.F. (1968). Corticofugal influence geniculate nucleus. Vis. Res. in the 8: 1285-1298. - 111 Jones, E.G. & Powell, T.S. (1969). An electron the mode of microscope study of termination of cortico-thalamic fibers within the sensory relay nuclei of the thalamus. Proc. Roy. Soc. B 172: 173-185. Kaas, J.H., Guillery, R.W. & Allman, J.M. (1972). Some principles organization in the dorsal LGN. ~~-•, Behav., Evol. Kalil, R.E. & Chase, R. (1970). Cortico-fugal influence of lateral geniculate neurons in the cat. of 6: 253-299. on activity J. Neurophys. 33: 459-474. Kawamura, S., Sprague, J.M. & Niimi, K. (1974). Corticofugal tions from the visual cortices to the thalamus, pretectum, and superior colliculus in the cat. J. Comp. Neurol. Lee, B.B., Creutzfeldt, O.D. & Elephant, A. (1979). The magno- projec- and parvo- cellular cells culate body to moving stimuli. 158: 339-362. reponses of the monkey's lateral geni- Exp. Br. Res. 35: 547-557. Leventhal, A.G. & Hirsch, H.B. (1978). Receptive field properties neurons in Neurophys. different lamina of of visual cortex of the cat. of J. 41: 948-962. Med. biol. Lund, J.S., Lund., R.D., Hendrickson, A.E., Bunt, A.H. & Fuchs, A.F. Levick, WQR. (1972). Another Engng. tungsten rnicroelec trode. 10: 510-5 15. (1975). The origin of efferent pathways from the primary visual cortex, area 17, of the macaque monkey as transport 287-304. of horseradish peroxidase. shown by retrograde l.• Comp. Neural. 164: - 112 Malcolm, L.J., Bruce, I.C. & Burke, w. (1970). lateral cat. Excitability of the geniculate nucleus in the alert, non-alert and sleeping ~-Br.Res. 10: 283-297. Movshon, J.A. (1975). The velocity tuning of single units in the striate cortex. J. Physiol. 249: 445-468. Nikara, T., Bishop, P.O. & Pettigrew, J.D. (1968). Analysis of nal correspondence by studying lateral geniculate K. reti- receptive fields of binocular single units in cat striate cortex. Noda, H., Tamaki, Y. & Iwama, cat (1972). Exp. Br. Res. Binocular nucleus of chronic cats. 6: 353-372. units in Br. Res. the 41: 81- 99. Pettigrew, J.D., Cooper, M.L. & Blasdel, G.G. (1979). Improved use of tapetal reflection for eye-position monitoring. Vis. Sci. Inv. Opthal. & 18: 490-495. Pettigrew, J.D., Nikara, T. & Bishop, P.O. (1968a). Responses to moving slits by single units in cat striate cortex. Exp. Br. Res. 6: 373-390. Pettigrew, J.D., interaction Nikara, T. & Bishop, P.O. (1968b). on single units in cat striate cortex: simultaneous stimulation by single correspondence. moving Exp. Br. Res. slit with receptive fields in the cat. (1977). Cells 6: 391-410. Pritchard, R.M. & Heron, w. (1960). Small eye movements of Canad. J. Psych. Ramachandran, Binocular 14: 131-137. v.s., Clarke, P.H. & Whitteridge, D. selective to binocular disparity in the cortex of newborn lambs. Nature 268: 333-335. - 113 - & Nikara, Rodieck, R.W., Pettigrew, J.D., Bishop, P.O. Residual cats. eye 7: 107-110. Rodieck, R.w. (1973). Inhibition in cat lateral role of Y system. geniculate Proc. Aust. Phys. Pharm. Soc. Rodieck, R.W. & Dreher, B. (1979). Visual Exp. Br. Res. Sanderson, K.J., Bishop, P.O. of the neurons. binocular Exp. Br. Res. Schmielau, F. interactions Br. Res. 120: 354-361. suppression & Darian-Smith, non- I. (1971). The proper- receptive fields of lateral geniculate 13: 178-207. in Br. Res. Bedworth, N. (1973). Inhibitory Y units in for the cat lateral genicula te nucleus. geniculate body of the cat. and from 35: 465-477. Singer, W. (1970). Inhibitory binocular interaction & 4: 167. Singer, w. (1977). The role of the visual cortex & binocular Singer, W. nucleus: eye in the lateral geniculate nucleus: a comparison of cat and monkey. ties (1967). movements in receptive field studies of paralysed Vis. Res. dominant T. in the lateral 18: 165-170. interaction between the cat lateral geniculate nucleus. X Br. Res. 49: 291-307. Singer~ w.~ Tretter, F. & Cynader, M. (1975). Organization of tive field properties J. Neurophys. with afferent and efferent connections. 38: 1080-1098. Singer, W. (1977). Control of thalamic transmission and Revs. ascending recep- reticular 57: 386-420. by corticofugal pathways in the visual system. Phys. - 114 - Stevens, J.K. & Gerstein, G.L. lateral geniculate neurons . (1976). Interactions .:I.• Neurophys. between 39: 239-256. Suzuki, H. & Kato, E. (1966). Binocular interaction at cat's geniculate body • .:I.• Neurophy~. cat lateral 29: 909-920. Suzuki, H. & Takahashi, M. (1970). Organization of lateral geniculate neurons in binocular inhibition. Br. Res. Tsumoto, T., Creutzfeldt, O.D. & Legendy, organization of the C.R. 23: 261-264. (1978). Functional corticofugal system from visual cortex to lateral geniculate nucleus in the cat. Exp. Br. Res. 32: 345- 364. Updyke, B.V. (1975). The patterns of projection of cortical areas 17, 18, and 19 onto the laminae of the dorsal lateral genfculate nucleus in the cat • .:I_. Comp. Neurol. Widen, L. & Marsan, C.A. (1960). Effects 163: 377-396. of corticopetal and cor- ticofugal impulses upon single elements of the dorsolateral geniculate nucleus. Exp. Neurol. 2: 468-502.