Relationship between laminar topology and retinotopy in the rhesus lateral geniculate nucleus: results from a functional atlas.

The primary focus of this paper is the abrupt transition that occurs midway through the rhesus lateral geniculate nucleus (LGN) from six layers posteriorly (conventionally numbered 1-6, ventral to dorsal) to four layers anteriorly. At this transition, layers 4 and 6 fuse into a single layer, as do layers 3 and 5, requiring an inversion of the stacking order of the cell categories making up layers 4 and 5. To understand the topology of this transition and its relationship to geniculate retinotopy, we have created a functional atlas of a rhesus LGN that affords three-dimensional views of morphology and retinotopy at a resolution of 25 microm. The projection of the path of the transition into visual space is highly biased toward lower visual fields, intersecting the upper vertical meridian at 6.4 degrees , the horizonal meridian at 15.4 degrees, and the lower vertical meridian at 25.0 degrees. Between inclinations of -31 degrees and 55 degrees, layers 3 and 5 merge through an elongated tear in layer 4 that subsumes the optic disk gap and extends medially and laterally; elsewhere, layers 4 and 6 merge through a tear in layer 5. These tears cause substantial violations of retinotopy and laminar integrity, so the inversion of layers 4 and 5 requires that the forces establishing retinotopy and grouping by cell class be locally overcome during morphogenesis. The transition and associated tears are evaluated in the context of recent computational models of geniculate morphogenesis. We have also used the atlas to estimate the borders of the binocular (55 approximately 62 degrees) and monocular (91 approximately 97 degrees) visual fields. Files containing the atlas are made publicly available on a website.

Reversible inactivation of subcortical sites by drug injection.

Reversible inactivation of subcortical targets by means of drug injections has been a powerful tool for revealing the contributions of discrete brain structures to behavior and the functional organization of the brain. This paper is intended to provide practical advice on this approach, including the choice of drug, means of delivering drugs, strategies for evaluating the action of injected drugs, and the application of this method to experiments with awake animals.

Effects of focal inactivation of dorsal or ventral layers of the lateral geniculate nucleus on cats' ability to see and fixate small targets.

To reveal contributions of different subdivisions of the lateral geniculate nucleus (LGN) to visuomotor behavior, segments of either layer A or the C layers were inactivated with microinjections of gamma-aminobutyric acid while cats made saccades to retinally stabilized spots of light placed either in affected regions of visual space or mirror-symmetric locations in the opposite hemifield. Inactivating layer A reduced the success rate for saccades to targets presented in affected locations from 82.4 to 26.8% while having no effect on saccades to the control hemifield. Saccades to affected sites had reduced accuracy and longer initiation latency and tended to be hypometric. In contrast, inactivating C layers did not affect performance. Data from all conditions fell along the same saccade velocity/amplitude function ("main sequence"), suggesting that LGN inactivations cause localization deficits, but do not interfere with saccade dynamics. Cerebral cortex is the only target of the A layers, so behavioral decrements caused by inactivating layer A must be related to changes in cortical activity. Inactivating layer A substantially reduces the activity of large subsets of corticotectal cells in areas 17 and 18, whereas few corticotectal cells depend on C layers for visually driven activity. The parallels between these behavioral and electrophysiological data along with the central role of the superior colliculus in saccadic eye movements suggests that the corticotectal pathway is involved in both deficits and remaining capacities resulting from blockade of layer A.

Effects of saccades on the activity of neurons in the cat lateral geniculate nucleus.

Effects of saccades on individual neurons in the cat lateral geniculate nucleus (LGN) were examined under two conditions: during spontaneous saccades in the dark and during stimulation by large, uniform flashes delivered at various times during and after rewarded saccades made to small visual targets. In the dark condition, a suppression of activity began 200-300 ms before saccade start, peaked approximately 100 ms before saccade start, and smoothly reversed to a facilitation of activity by saccade end. The facilitation peaked 70-130 ms after saccade end and decayed during the next several hundred milliseconds. The latency of the facilitation was related inversely to saccade velocity, reaching a minimum for saccades with peak velocity >70-80 degrees /s. Effects of saccades on visually evoked activity were remarkably similar: a facilitation began at saccade end and peaked 50-100 ms later. When matched for saccade velocity, the time courses and magnitudes of postsaccadic facilitation for activity in the dark and during visual stimulation were identical. The presaccadic suppression observed in the dark condition was similar for X and Y cells, whereas the postsaccadic facilitation was substantially stronger for X cells, both in the dark and for visually evoked responses. This saccade-related regulation of geniculate transmission appears to be independent of the conditions under which the saccade is evoked or the state of retinal input to the LGN. The change in activity from presaccadic suppression to postsaccadic facilitation amounted to an increase in gain of geniculate transmission of approximately 30%. This may promote rapid central registration of visual inputs by increasing the temporal contrast between activity evoked by an image near the end of a fixation and that evoked by the image immediately after a saccade.

Thalamic control of cat lateral suprasylvian visual area: relation to patchy association projections from area 18.

In this study, we examined functional contributions of major subdivisions of the lateral geniculate nucleus to the cat's lateral suprasylvian visual area (LS) in relation to the patchy horizontal distributions of association inputs. Multiple-unit activity driven via the contralateral eye was assessed during reversible blockade of the retinotopically corresponding part of layer A, the C layers as a group, or the medial interlaminar nucleus (MIN). Inactivating each of these targets reduced activity at some cortical sites, with inactivation of layer A having, on average, the largest effect. Activity was rarely abolished by inactivation of a single target, indicating that most LS sites receive multiple inputs. Dependence on layer A was strongly correlated with the horizontal distribution of association inputs from area 18. Closely spaced injections of anatomical tracers into extensive regions of area 18 resulted in patches of terminal label in lateral suprasylvian cortex. Activity inside the patches was relatively dependent on layer A, whereas that outside the patches was not. Dependence on the MIN and layer A were negatively correlated, suggesting that inputs dominated by the MIN and layer A were concentrated in independent sets of patches. These results indicate that the anatomically observed patchy projections reflect the functional consequences of geniculate lamination. The A layers are high-acuity relays, whereas the MIN is probably a specialization for dim-light vision (Lee et al., 1984; Lee et al., 1992). We propose that the partial overlap of inputs dominated by the A layers and the MIN allows dynamic shifts in their relative contributions to LS responses, optimizing the balance of high-acuity and high-sensitivity channels over a wide range of illumination conditions.

Thalamic control of cat area-18 supragranular layers: simple cells, complex cells, and cells projecting to the lateral suprasylvian visual area.

The goal of this study was to determine the effects of inactivating layer A or the C layers of the cat lateral geniculate nucleus on the supragranular layers of area 18, including cells antidromically activated from the lateral suprasylvian visual area (LS). Isolated cells were visually driven via the contralateral eye while the retinotopically corresponding regions of layer A or, in some cases, the C layers were reversibly inactivated with injections of cobaltous chloride. Simple cells were frequently encountered and were on average more dependent upon layer A than were complex cells, a result qualitatively similar to that found previously in area 17 (Malpeli, 1983; Malpeli et al., 1986). However, the influence of the C layers on area 18 was much more apparent than for area 17. In area 18, as in area 17, the dependence of simple cells on particular geniculate layers appears to follow the terminal patterns of the major direct geniculate inputs. Those simple cells most dependent on layer A were located in lower layer 3. Simple cells in upper layer 3, like complex cells, showed little dependence on layer A, but were strongly dependent upon the C layers. All cells antidromically activated from LS were simple cells with rapidly conducting axons. They had, on average, the same moderately strong dependence on layer A as the patches of LS receiving area 18 input (Lee et al., 1997), supporting the conclusion that the influence of layer A in these patches is largely transmitted via association inputs from area 18. These results demonstrate that simple cells play a major role in association pathways.

Measuring eye position with the double magnetic induction method.

The double magnetic induction method of monitoring eye movements eliminates problems associated with the external wire leads of the standard magnetic search-coil technique by substituting a closed metal ring for the eye coil. However, external coils used in the double-induction method must be adjusted at the beginning of each session, and the system is very sensitive to mechanical disturbances. An apparatus is described whose ease of adjustment and good mechanical stability make the double-induction method as convenient and reliable as the standard method. The performance of the system is evaluated for different physical arrangements of critical components, and some practical advice on various aspects of the double-induction method is offered.

Laminar and retinotopic organization of the macaque lateral geniculate nucleus: magnocellular and parvocellular magnification functions.

The laminar morphology and electrophysiologically determined retinotopic organization of a single rhesus macaque lateral geniculate nucleus (LGN) were reconstructed on series of coronal, sagittal, and horizontal cuts through a three-dimensional computer representation of the nucleus. Neurons were counted in this same nucleus, allowing the magnification functions (cells/degree2 as functions of eccentricity) of magnocellular and parvocellular layers to be determined after eliminating the effects of nonuniform volume shrinkage. Parvocellular magnification was approximately 10,000 times higher in the foveola than in the far periphery. On average, magnocellular neurons made up 2.6% of the LGN in the central 2 degrees (but probably a smaller fraction in the central fovea). The magnocellular portion increased steadily with eccentricity to 27% in the far periphery. Thus the magno/parvo ratio increases from foveola to far periphery by a factor of at least 14. The parvocellular magnification function matches estimates of cortical magnification, whereas the density of magnocellular afferents to cortex increases monotonically with eccentricity. At the posterior pole of the nucleus, the numbers of layers are reduced through a fusion of two layers and the disappearance of one or two others, a feature that may be associated with the foveal ipsilateral hemifield representation.

Global form and singularity: modeling the blind spot's role in lateral geniculate morphogenesis.

Optic nerve terminals segregate by functional class into distinct layers in the lateral geniculate nucleus, the thalamic relay nucleus of the visual system. In the rhesus monkey, the number of geniculate layers changes abruptly from six posteriorly (central vision) to four anteriorly (peripheral vision). The plane of transition between these patterns passes through small laminar gaps corresponding to the perceptual blind spot caused by the exit of the optic nerve from the eyeball. However, this plane of transition has no apparent functional link to the blind spot. A thermodynamic model of geniculate morphogenesis supports the hypothesis that the blind spot traps the transition in its stereotypic position by introducing a singularity in an otherwise smooth gradient in forces guiding the development of geniculate morphogenesis. This relation suggests that small-scale anomalies may be important in the determination of large-scale patterns in biological structure.

Responses of neurons in primary visual cortex are modulated by eye position.

1. We tested the effects of eye position on the visual excitability of 88 neurons in the primary visual cortex of awake cats trained in oculomotor tasks. For most cells, we examined responses evoked by retinotopically identical stimuli for centered gaze, 8 degrees to the left of center, and 8 degrees to the right of center. 2. An effect of eye position was observed for 40% of the cells. For 13%, responsiveness varied by a factor of 2 or more. Most commonly, response was maximal with gaze shifted to one side, minimal when shifted to the opposite side, and intermediate for centered fixation. The exceptions were four cells for which excitability varied symmetrically with fixations to either side of center. 3. Variability in excitability associated with eye position is a wide-spread phenomenon, having been observed in the lateral geniculate nucleus, V1, and extrastriate cortex. These results are consistent with the belief that such variability is utilized in constructing a head-centered frame of reference from a retinotopic input.

Acuity-sensitivity trade-offs of X and Y cells in the cat lateral geniculate complex: role of the medial interlaminar nucleus in scotopic vision.

1. The cat medial interlaminar nucleus (MIN) receives inputs almost exclusively from tapetal retina, suggesting that the MIN has a special role in dim-light vision. In this study we compared the sensitivities of cells in the MIN with those in layers A and magnocellular C of the lateral geniculate nucleus (LGNd), using drifting sinusoidal gratings to determine contrast thresholds as a function of spatial frequency and retinal adaptation level over the entire scotopic range. 2. About one-half of the cells recorded in the MIN and layer A had brisk responses that could be nulled by properly positioned, counterphased sinusoidal gratings, and were classified as X cells. The rest of the cells in the MIN and layer A, as well as all cells recorded in layer C, were Y cells. 3. MIN cells had higher contrast sensitivity than layer A cells for low spatial frequencies (0.15 cycles/deg and below) over a wide range of adaptation levels, both overall and for separate comparisons within X or Y cells. Layer C Y cells were intermediate in sensitivity between MIN and layer A Y cells. For low spatial frequencies, Y cells as a group were more sensitive than X cells, whereas the reverse was true for high spatial frequencies. 4. These data enable one to determine the lowest adaptation level at which stimuli of a given contrast can be detected for a given structure. At the lowest spatial frequencies, the MIN can function at adaptation levels approximately 1 log unit below layer A, averaged over all stimulus contrasts. In contrast, the tapetum lowers luminance threshold by at most 0.16 log unit. 5. For scotopic conditions and eccentricities within 15 degrees of the area centralis, contrast sensitivity decreases with eccentricity for low spatial frequencies and remains flat or slightly increases for high spatial frequencies. This relationship, which is opposite to that found for photopic vision, is strongest for MIN Y cells. 6. These data support the hypothesis that the retinal conflict between sensitivity and acuity is ameliorated in the CNS through separate thalamic relays with different degrees of afferent convergence. MIN cells have higher luminance sensitivity than layer A cells, but at the expense of acuity. Layer C appears to occupy an intermediate position in this trade-off.

A new method of mounting and directing chronically implanted microdrives.

A new method of mounting a microdrive on the skull and adjusting the trajectories of microelectrodes is described. The key to this system is a swiveling guide tube held in a small, skull-mounted base by a low-melting-point metal alloy. The microelectrode is advanced via a modified, commercially available, miniature microdrive screwed onto the guide tube. To change the trajectory of the electrode, the alloy is melted in place, and the guide tube swiveled to a different angle. The microelectrode or the trajectory of the pass may be changed in a few minutes while easily maintaining aseptic conditions. The entire assembly is small enough so that several can be simultaneously implanted on the skull of a cat. A metal crown is used to fix the head during recording sessions or to hold a fiberglass cap that protects all skull implants between sessions. Since the microdrives need not be removed between recording sessions, an electrode pass may be continued from day to day. Although guide tubes are generally employed only for subcortical structures, this arrangement also works well when recording from cortex in deeper parts of gyri.

Area 18 corticotectal cells: response properties and identification of sustaining geniculate inputs.

1. We examined the response properties and geniculate inputs of 35 antidromically identified corticotectal (CT) cells within area 18 of the paralyzed, anesthetized cat. Twenty-three were either standard complex or hypercomplex, 11 were special complex, and 1 was simple. 2. The response properties of CT cells in area 18 were in general quite similar to those examined in a previous study of area 17 CT cells, including similar proportions of standard and special complex CT cells, virtually identical length-response functions, and similar orientation and direction tuning. 3. Area 18 CT cells are rapidly conducting. They are considerably faster than area 17 CT cells. 4. We investigated the composition of thalamic inputs to CT cells by reversibly inactivating a portion of layer A and/or the C layers of the dorsal lateral geniculate nucleus with injections of cobaltous chloride. Blocking layer A strongly attenuated the visual responsiveness of about half of the cells tested. Blocking the C layers alone generally had only moderate effects, but simultaneous blockade of layer A and the C layers demonstrated a substantial C-layer input to many cells. Unlike area 17 in which there is a strong correlation between CT cell class and dependence on layer A, no single receptive-field parameter nor set of parameters was correlated with dependence on layer A. However, cells least affected by simultaneous blockade of layer A and the C layers were special complex, suggesting that, as in area 17, area 18 special complex CT cells integrate more geniculate inputs than standard complex CT cells. 5. We propose that the similarities of response properties of area 17 and area 18 CT cells results from their participation in similar interlaminar columnar circuits and that differences in the patterns of geniculate control reflect differences in the global patterns of geniculate inputs to these two areas.

Paths of information flow through visual cortex.

The main route of information flow in the cerebral cortex is from the middle layers of cortex to upper and lower layers. However, upper layers of the cat primary visual cortex can be directly driven by inputs from secondary visual cortex when activity in middle layers is disrupted. Upper-layer activity can be driven either by middle layers or by direct corticocortical inputs. One consequence of this result is that areas of cortex thought to be carrying out low-order analysis may be able to extract considerable information from higher order areas.

Cat area 17. I. Pattern of thalamic control of cortical layers.

Reversible inactivation of individual layers of the cat lateral geniculate and medial interlaminar nuclei was used to investigate the necessary and sufficient inputs for maintaining visually driven activity and receptive field properties in area 17. Neither orientation selectivity nor direction selectivity depends on any individual geniculate layer. We identified two groups of cortical layers on the basis of the pattern of thalamic inputs providing visual driving through the contralateral eye. One group, consisting of layers 4 and 6, has geniculate layer A as its only necessary and sufficient input. The other, consisting of supragranular layers, integrates at least two sufficient thalamic inputs, one of which is layer A. Several major receptive field properties are independently generated in these two groups of layers.

Cat area 17. II. Response properties of infragranular layer neurons in the absence of supragranular layer activity.

Response properties of cells in the infragranular layers of cortical area 17 of the cat were examined in the absence of input from supragranular layers. Supragranular activity was silenced either reversibly by cooling the surface of cortex or permanently by making a cryogenic lesion of the supragranular layers. Visually driven responses of cells throughout the cortical column were recorded with a linear array of electrodes. Most infragranular layer cells continued to be visually responsive in the absence of supragranular layer input. These cells were similar to normal infragranular layer cells on measures of visual responsiveness, orientation selectivity, and direction selectivity. Special complex, but not standard complex, cells were absent in layer 5 when supragranular layers were destroyed. We found no evidence for a selective effect of removal of supragranular activity on the response properties of cells in layer 6. We propose that the intracolumnar projection from the supragranular layers drives the special complex cells of layer 5, but is not necessary for the visual driving of most other infragranular layer cells. This projection does not impose selectivity for stimulus orientation or direction on the remaining active cells of the infragranular layers.

Cat area 17. III. Response properties and orientation anisotropies of corticotectal cells.

The receptive field properties of antidromically identified corticotectal (CT) cells in area 17 were explored in the paralyzed, anesthetized cat. To compare these with another population of infragranular cells, we also examined the receptive field properties of cells in layer 6. Sixty percent of our sample of CT cells showed increased response to increased stimulus length (length summation) and were classified as standard complex cells. The other 40% showed little or no length summation, were generally end stopped, and were classified as special complex cells. Standard and special complex CT cells have complementary orientation anisotropies: the distribution of orientation preferences of standard complex cells is biased toward obliquely oriented stimuli, whereas special complex cells are biased toward horizontally and vertically oriented stimuli. The receptive fields of the cells in our sample were primarily along the horizontal meridian so we cannot determine if these anisotropies are defined relative to the vertical meridian or relative to the meridian passing through the receptive field. The effects of these anisotropies in preferred orientation are minimized by the broad orientation tuning of CT cells. There was no simple relationship between the direction bias of CT cells and the reported direction bias of tectal cells. In contrast to the heterogeneity of corticotectal cells, layer 6 cells uniformly showed strong length summation, tight orientation tuning, and little spontaneous activity.

Cat area 17. IV. Two types of corticotectal cells defined by controlling geniculate inputs.

The dependence of cat area 17 corticotectal (CT) cells on specific subdivisions of the dorsal lateral geniculate (LGN) and medial interlaminar nuclei (MIN) was examined using reversible inactivation techniques. Inactivation of layer C of the LGN or layer 1 of the MIN did not block visual activity of CT cells driven through the contralateral eye. Inactivation of layer A of the LGN revealed two populations of CT cells: one strongly dependent on layer A and one whose visually driven activity survived layer A inactivation. CT cells that responded best to short stimuli (special complex cells) were least dependent on layer A, whereas cells that responded best to long stimuli (standard complex cells) were most dependent on layer A. We propose a model of the intracortical circuitry of these two types of CT cells. Standard complex cells, which are heavily dependent on layer A, receive sustaining visual input through layers 4 and/or 6. Special complex cells, which are not dependent on any single layer of the lateral geniculate nucleus, receive sustaining visual input from supragranular layers.

Somata-selective lesions induced by cobaltous chloride: a parametric study.

Repeated injections of cobaltous chloride destroy neuronal somata while sparing fibers-of-passage. This effect is dose-dependent and excess cobalt destroys fibers-of-passage as well as somata. By manipulating the concentration and volume of cobaltous chloride injected into rat lateral geniculate nuclei, we defined a range of parameters producing highly selective destruction of somata for regions 1 mm in diameter or smaller.

Cat medial interlaminar nucleus: retinotopy, relation to tapetum and implications for scotopic vision.

The medial interlaminar nucleus (MIN) of the cat was electrophysiologically mapped in sufficient detail to resolve individual laminae and to allow reconstruction of isoazimuth and isoelevation lines in coronal, sagittal, and horizontal planes. The electrophysiologically defined laminar pattern was in agreement with that revealed anatomically in the same animal, as well as with the general pattern revealed by anatomical methods in several unmapped nuclei. The MIN is made up of three distinct layers, each receiving inputs from one hemi-retina but none representing an entire hemifield. We confirm the findings of Guillery et al. (19) that the contralateral hemifield is represented in layers 1 and 2 through the contralateral and ipsilateral eyes, respectively, and that layer 3 represents the ipsilateral hemifield through the contralateral eye. Elevation and absolute value of azimuth are represented continuously through the MIN. When an isoazimuth line crosses the border between layer 3 and either layer 1 or 2, the absolute value of azimuth is maintained but the sign of the azimuth changes. Adjacent points on either side of this border represent mirror symmetrical visual directions on opposite sides of the vertical meridian. This indicates that the distance from the vertical meridian is an independently coded parameter within the geniculate complex. There is virtually no nasotemporal overlap in any layer of the MIN. The function relating magnification (mm3 per steradian) to eccentricity is strikingly similar to the function relating retinal ganglion cell density to eccentricity, suggesting that a constant fraction of retinal ganglion cells project to the MIN at all eccentricities. Most of the volume of each MIN layer is devoted to lower visual fields. Analysis of the geniculate retinotopic maps of Sanderson (46) reveals no equivalent bias toward lower visual fields in the dorsal lateral geniculate nucleus. The MIN represents a region of retina roughly coincident with the tapetum, suggesting a role of the MIN in dim-light vision.