Photically sustained on-responses of units in posterior hippocampal gyrus of awake monkey

Representation of Contralateral Visual Space in the Human Hippocampus

The initial encoding of visual information primarily from the contralateral visual field is a fundamental organizing principle of the primate visual system. Recently, the presence of such retinotopic sensitivity has been shown to extend well beyond early visual cortex to regions not historically considered retinotopically sensitive. In particular, human scene-selective regions in parahippocampal and medial parietal cortex exhibit prominent biases for the contralateral visual field. Here, we used fMRI to test the hypothesis that the human hippocampus, which is thought to be anatomically connected with these scene-selective regions, would also exhibit a biased representation of contralateral visual space. First, population receptive field (pRF) mapping with scene stimuli revealed strong biases for the contralateral visual field in bilateral hippocampus. Second, the distribution of retinotopic sensitivity suggested a more prominent representation in anterior medial portions of the hippocampus. Finally, the contralateral bias was confirmed in independent data taken from the Human Connectome Project (HCP) initiative. The presence of contralateral biases in the hippocampus, a structure considered by many as the apex of the visual hierarchy, highlights the truly pervasive influence of retinotopy. Moreover, this finding has important implications for understanding how visual information relates to the allocentric global spatial representations known to be encoded therein.SIGNIFICANCE STATEMENT Retinotopic encoding of visual information is an organizing principle of visual cortex. Recent work demonstrates this sensitivity in structures far beyond early visual cortex, including those anatomically connected to the hippocampus. Here, using population receptive field (pRF) modeling in two independent sets of data we demonstrate a consistent bias for the contralateral visual field in bilateral hippocampus. Such a bias highlights the truly pervasive influence of retinotopy, with important implications for understanding how the presence of retinotopy relates to more allocentric spatial representations.

Computer simulation study of the relationship between the profile of excitatory postsynaptic potential and stimulus-correlated motoneuron firing

This paper shows the results of computer simulation of changes in motoneuron (MN) firing evoked by a repetitively applied synaptic volley that consists of a single excitatory postsynaptic potential (EPSP). Spike trains produced by the threshold-crossing MN model were analyzed as experimental results. Various output functions were applied for analysis; the most useful was a peristimulus time histogram, a special modification of a raster plot and a peristimulus time frequencygram (PSTF). It has been shown that all functions complement each other in distinguishing between the genuine results evoked by the excitatory volley and the secondary results of the EPSP-evoked synchronization. The EPSP rising edge was best reproduced by the PSTF. However, whereas the EPSP rise time could be estimated quite accurately, especially for high EPSP amplitudes at high MN firing rates, the EPSP amplitude estimate was also influenced by factors unrelated to the synaptic volley, such as the afterhyperpolarization duration of the MN or the amplitude of synaptic noise, which cannot be directly assessed in human experiments. Thus, the attempts to scale any estimate of the EPSP amplitude in millivolts appear to be useless. The decaying phase of the EPSP cannot be reproduced accurately by any of the functions. For the short EPSPs, it is extinguished by the generation of an action potential and a subsequent decrease in the MN excitability. For longer EPSPs, it is inseparable from the secondary effects of synchronization. Thus, the methods aimed at extracting information about long-lasting and complex postsynaptic potentials from stimulus-correlated MN firing, should be refined, and the theoretical considerations checked in computer simulations.

A neuromagnetic view of hippocampal memory functions

Bursts of highly synchronized discharges of 4-7Hz sinusoidal wave activity can be recorded from the hippocampus during rapid eye movement (REM) sleep. These rhythmic discharges, the hippocampal theta activity, are generated in the dentate granule cells and the pyramidal cell layers of the CA1 field of the hippocampus. The physiological function of the hippocampal theta activity is elusive. The occurrence of this rhythm throughout the REM sleep stage suggests that it is related to some fundamental neurophysiological phenomena associated with REM sleep, particularly consolidation of memory processes. Synchronous oscillations among a population of neurons are expected to yield stronger, more coherent associated magnetic fields which, through their influence back on the electrical fields via induction, would exert an independent effect on the electrical activity of hippocampal neurons and additionally, could foster and reinforce these oscillations through self-induction. The snail-shaped structure of the hippocampal formation, which resembles a solenoid embedded in the temporal lobe, would be expected to amplify these magnetic fields. Additionally, the discovery of large ferromagnetic particles in the human hippocampus suggests that it may function as a large iron-core electromagnet. It is proposed that memory traces may be encoded or decoded magnetically and analogous to a videotape, each encoding unit (i.e., synapse, set of synapses or glial cell) could be magnetized in one direction, or the other through the flow of a strong, AC magnetic field along the hippocampal formation. The encoding of memory traces in the hippocampal formation may ultimately reflect an electromagnetic phenomenon.

Functional comparison of neuronal properties in the primate posterior hippocampus and parahippocampus (area TF/TH) during different behavioural paradigms involving memory and selective attention

Monkeys were trained on a delayed match-to-sample (DMS) task. In addition a standardized behavioural trial was performed which involved an experimenter approaching the animal in certain sequence and presenting edible or other objects ('raisin trial'). Neuronal activity of 617 units was recorded in the posterior parahippocampus (PH) and in the posterior hippocampus (H). In many cases, we compared the activity of the same neuron in different tasks. 32.7% of the 455 PH neurons and 28.5% of the 130 H cells responded during the presentation of the visual stimuli in the DMS task. These responses were only mildly influenced by the physical dimensions of the visual stimulus, but often depended on the context in which the stimuli were presented. There was no differential response to the second stimulus that clearly depended on the nature of the first stimulus. 6.2% of the PH units, but none in H, responded in relation to the reward. 4.4% of the PH neurons, but none in H, showed a mild response during the interstimulus interval. 38.1% of 215 PH neurons and 37.8% of 45 H cells responded during one or more phases of the raisin trial. These responses were not related to the physical dimensions of the sensory stimuli. 210 PH and 41 H units were investigated during the DMS task as well as during the raisin trial. 18.1% (PH) and 12.2% (H) of the units responded during the DMS task, but not during the raisin trial; 17.1% (PH) and 36.6% (H) responded vice versa. A response in both trials was found in 17.1% of the PH neurons, but in none of the H cells. There were also other PH unit types showing responses during different aspects of the DMS task and even in other control paradigms, while no such overlap was encountered in H. Our results suggest a function of H and PH in the evaluation of the behavioural significance of sensory information. It may be this aspect which leads to anterograde memory disturbances after lesion of these areas. Since representation of neuronal information was found to be more specific in H, a possible function as an 'evaluation index' is discussed.

Focal medial temporal lobe spike-wave complexes evoked by a memory task

Several factors are known to modulate the occurrence of interictal spike-wave complexes (SWC). During performance of a recent memory task by a patient with uncontrolled complex partial seizures and implanted electrodes, visually presented stimuli consistently evoked focal medial temporal lobe interictal SWC with an average latency of 228 +/- 29 ms. The SWC were more often lateralized to the left medial temporal lobe. No consistent change in interictal spike rate was observed during performance of several other cognitive tasks during which visual or auditory stimuli were presented. This suggests that performance of specific cognitive tasks--in this case, recognition memory of words--can trigger interictal SWC.

Cingulate cortex of the rhesus monkey: I. Cytoarchitecture and thalamic afferents

The cytoarchitecture and thalamic afferents of cingulate cortex were evaluated in the rhesus monkey (Macaca mulatta). Area 24 has three divisions of which area 24a is adjacent to the callosal sulcus and has the least laminar differentiation. Area 24b has more clearly defined layers II, III, and Va, and area 24c, which forms the lower bank of the anterior cingulate sulcus, has a particularly dense layer III. Area 23 also has three divisions, each of which has a distinct layer IV. Area 23a is adjacent to the callosal sulcus and has the thinnest layers II-IV, which have the same cell density as layers V and VI. Area 23b has the largest pyramids in layers IIIc and Va, and area 23c, in the depths of the posterior cingulate sulcus, has the broadest external and thinnest internal pyramidal layers. Finally, areas 29 and 30 are located in the posterior depths of the callosal sulcus. Two divisions of area 29 are apparent: one with a granular layer directly adjacent to layer I (area 29a-c) and another with differentiation of layers III and IV (area 29d). Area 30 has a dysgranular layer IV. Injections of the retrograde tracer horseradish peroxidase (HRP) were made into subdivisions of cingulate cortex in the monkey. Area 25 received thalamic input mainly from the midline parataenial (Pt), central densocellular (Cdc), and reuniens nuclei as well as from the dorsal parvicellular division of the mediodorsal nucleus (MDpc). A less dense projection also originated in the intralaminar parafascicular (Pf), central superior, and limitans (Li) nuclei as well as the medial division of the anterior nuclei (AM). Areas 24a and 24b received most thalamic afferents from fusiform and multipolar cells in the Cdc and Pf nuclei with fewer from the ventral anterior (VA) and MDpc and MD densocellular (MDdc) nuclei and only minor input from AM. Most input to premotor cingulate area 24c appeared to originate in VA, MDdc, and Li. Area 29 received the most dense input from nuclei traditionally associated with limbic cortex including the anteroventral (AV), anterodorsal (AD), and laterodorsal (LD) nuclei. Areas 23a and 23b, in contrast, did not receive AV, AD, or LD input, but the greatest proportion of their thalamic afferents arose in AM. Less-pronounced input also came from the lateroposterior (LP), medial pulvinar, and MDdc nuclei.(ABSTRACT TRUNCATED AT 400 WORDS)

Cingulate cortex of the rhesus monkey: II. Cortical afferents

Cortical projections to subdivisions of the cingulate cortex in the rhesus monkey were analyzed with horseradish peroxidase and tritiated amino acid tracers. These projections were evaluated in terms of an expanded cytoarchitectural scheme in which areas 24 and 23 were divided into three ventrodorsal parts, i.e., areas 24a-c and 23a-c. Most cortical input to area 25 originated in the frontal lobe in lateral areas 46 and 9 and orbitofrontal areas 11 and 14. Area 25 also received afferents from cingulate areas 24b, 24c, and 23b, from rostral auditory association areas TS2 and TS3, from the subiculum and CA1 sector of the hippocampus, and from the lateral and accessory basal nuclei of the amygdala (LB and AB, respectively). Areas 24a and 24b received afferents from areas 25 and 23b of cingulate cortex, but most were from frontal and temporal cortices. These included the following areas: frontal areas 9, 11, 12, 13, and 46; temporal polar area TG as well as LB and AB; superior temporal sulcus area TPO; agranular insular cortex; posterior parahippocampal cortex including areas TF, TL, and TH and the subiculum. Autoradiographic cases indicated that area 24c received input from the insula, parietal areas PG and PGm, area TG of the temporal pole, and frontal areas 12 and 46. Additionally, caudal area 24 was the recipient of area PG input but not amygdalar afferents. It was also the primary site of areas TF, TL, and TH projections. The following projections were observed both to and within posterior cingulate cortex. Area 29a-c received inputs from area 46 of the frontal lobe and the subiculum and in turn it projected to area 30. Area 30 had afferents from the posterior parietal cortex (area Opt) and temporal area TF. Areas 23a and 23b received inputs mainly from frontal areas 46, 9, 11, and 14, parietal areas Opt and PGm, area TPO of superior temporal cortex, and areas TH, TL, and TF. Anterior cingulate areas 24a and 24b and posterior areas 29d and 30 projected to area 23. Finally, a rostromedial part of visual association area 19 also projected to area 23. The origin and termination of these connections were expressed in a number of different laminar patterns. Most corticocortical connections arose in layer III and to a lesser extent layer V, while others, e.g., those from the cortex of the superior temporal sulcus, had an equal density of cells in both layers III and V.(ABSTRACT TRUNCATED AT 400 WORDS)

Habituation of human limbic neuronal response to sensory stimulation

Hippocampal, parahippocampal gyrus, and amygdalar neuronal responses to visual and acoustic stimuli were analyzed during trains of several hundred stimulus repetitions as part of an investigation of sensory pathways to medial temporal lobe structures in complex-partial epilepsy patients who were being monitored with depth electrodes. Ten percent of more than 500 single and multiple units tested were responsive to simple sensory stimuli. The majority of the responsive units were recorded in the posterior parahippocampal gyrus (HG) during visual stimulation. Although neurons in pes hippocampi (PH; Ammons's horn) were also responsive to photic stimuli, no visually responsive units were found in amygdala. Very few units were responsive to acoustic stimuli, and these were found only in PH and amygdala, and not in HG. Significant trends of increase or decrease in response amplitude during trains of stimuli were found in all acoustically responsive units. Significant trends of visual response amplitude increase or decrease were found in 20% of PH units, and in 44% of HG units. Mean latencies of acoustically responsive units were longer than those of visually responsive units, and latencies of PH sensory units showing decremental response were longer than nondecremental PH units. Rate of response decrement was usually linear for acoustic responses and exponential for visual responses. The response dynamics of medial temporal lobe neurons are compared with those described in the animal limbic system and are related to habituation of human sensory evoked scalp potentials.

Hippocampal influence on amygdala unit activity in awake squirrel monkeys

The influence of hippocampal electrical stimulation on amygdala extracellular unit activity was studied in the awake squirrel monkey. Hippocampal influence was topographically organized. Stimulation elicited responses in 20% of 476 units tested. Ipsilateral anterior stimulation was more effective than posterior or contralateral stimulation. In the 6 areas with more than 25 tested units, the basomedial nucleus had the highest percentage of responsive units (39%), followed by the accessory basolateral (33%), central (22%y, basolateral (21%), and lateral (5%) nuclei, and the anterior amygdala area (4%). Initial excitation (E) was more prevalent than initial inhibition (I) in the central (90% E vs 10% I) and basomedial (82% E vs 18% I) nuclei; but initial inhibition was more common in basolateral (37% E vs 63% I) and accessory basolateral (33% E vs 67% I) nuclei. The mean response latency was 30.8 msec, ranging from 12 to 130 msec. The basomedial nucleus appears to receive the most potent hippocampal influence. Response characteristics are consistent with a hypothesized relay of nonfornix hippocampal influences on basal forebrain and hypothalamus via the basomedial nucleus. Possible pathways of hippocampal influence and the implications of our results for concepts of hippocampal and amygdala function are discussed.

An olfactory input to the hippocampus of the cat: field potential analysis

Hippocampal responses to electrical stimulation of the prepyriform cortex in the cat were studied both in acute experiments under halothane anesthesia and in awake cats with chronically indwelling electrodes. Analysis of field potentials and unit activity indicated the extent to which different hippocampal subareas were activated, the laminar level at which the synaptic action took place and the dynamics of the evoked responses. It was found that: (1) the main generator of evoked responses in the hippocampus upon prepyriform cortex stimulation is localized in the fascia dentata and CA3 (CA1 pyramidal cells, and probably also subiculum cells, are activated but in a lesser degree); (2) the initial synaptic activity takes place at the most distal part of the dendrites of fascia dentata granuhat leads to a transient increase in the firing rate of the hippocampal units, which is often followed by a long-lasting decrease in firing rate. We conclude that the pathway from the prepyriform cortex via lateral entorhinal cortex to hippocampal neurons may enable olfactory inputs to effectively excite hippocampal neurons.

Pulvinar and lateral geniculate neuronal activity in the cat during operantly conditioned appetitive behavior

Cats were trained to press a lever for 0.5--1.0 ml of milk reward both in the presence and absence of ambient light. 'Floating' microelectrodes were implanted in the pulvinar and lateral geniculate nuclei to allow chronic recording from single or multiple neurons over a period of 1--16 days. Thirty-one of 36 pulvinar neurons (86%) showed a consistent periodicity in the firing rate that was clearly related to certain phases of operant behavior. In the presence of ambient light, the firing rate was maximal shortly prior to or immediately after a rewarded bar press, particularly during the animal's initial contact with reward (the first lap). This burst of activity was followed by an abrupt and strong inhibition of firing which was, in most instances, associated with high voltage slow wave (6--9 c/sec) postreinforcement synchronization (PRS) of the electroencephalogram over the visual and association cortices as well as in the pulvinar. Postreinforcement inhibition of neuronal activity was also observed in the lateral geniculate. Neuronal firing in both nuclei dropped significantly below the respective mean rate observed for each nucleus during relaxed wakefulness. In the dark, however, the reward-induced inhibition was abolished in both nuclei. Moreover, some pulvinar neurons in the dark showed a conspicuous and significant increase in firing during the consummatory response (dark reversal pattern). These findings suggest that a transient but powerful inhibitory process triggered by positive reinforcement depends on visual input. Firing patterns of some pulvinar neurons were noted to be entrained not only by the rhythm of PRS oscillations (inhibitory phasing and postinhibitory discharges) but also by the much slower 3--4 c/sec rhythm of lapping and licking. Furthermore, the patterns of entrainment by initial lapping appear to be different from those observed during the later phase of the consummatory response, thus indicating that the pulvinar may receive information regarding the content of the delivery cup. No entrainment of neuronal activity was observed in the lateral geniculate. These findings reveal remarkable plasticity in the pulvinar neurons that is consistent with their role in polysensory integration.

Dependence of reward contingent positive variation (RCPV) and cortical synchronization on visual input in the cat

Cats trained to press a lever for milk reward display high amplitude low frequency electrocorticographic post-reinforcement synchronization (PRS) associated with epicortical steady potential shift, termed Reward Contingent Positive Variation (RCPV), over the parieto-occipital cortex. Previously, it was found that the PRS-RCPV phenomenon depends on unpatterned light input devoid of any conceivable information or conditional property. Although training in the dark most likely eliminated the "novelty" factor and the resulting orienting reaction to "light off" as a cause of PRS-RCPV suppression, a possibility could not be excluded that the absence of light had a negatively reinforcing property preventing the emergence of PRS-RCPV. Hence, in the present study an experimental paradigm was used in which "light off" and/or "tone on" cues signalled the availability of reward, the "light off" periods extending through the consummatory response. Seven out of 9 cats, despite 2-4 months of daily training, and 85% correct timing of bar press performance, never showed a PRS-RCPV response (one animal showed only sporadic and poorly developed patterns.) However, when the light was turned on during consumption, in 85% of such tests a fully developed PRS-RCPV emerged in a stimulus-bound manner. One animal was an exception and showed PRS-RCPV both in the absence and presence of light. It was concluded that light input plays a crucial role in the emergence of the PRS-RCPV phenomenon and therefore also participates in the integration of gustatory input.

Monkey retinal ganglion cells: morphometric analysis and tracing of axonal projections, with a consideration of the peroxidase technique

This paper presents evidence on the retinal distribution and central projections of retinal ganglion cells of various cell body sizes in the adult macaque monkey. The ganglion cell sizes have been determined by computer assisted measurement of camera lucida drawings at various eccentricities of both flat mounted and sectioned retinae. The pattern of projections of individual ganglion cells to the dorsal lateral geniculate nucleus and superior colliculus has been studied using retrograde axonal transport of horseradish peroxidase. Following peroxidase injections into the parvocellular laminae of the geniculate, virtually every ganglion cell was labeled within a circumscribed zone of the retina known to project to the region of the geniculate immediately surrounding the injection needle tip. After magnocellular injections, only the largest cells of the peripheral retina and approximately 26% of the ganglion cells of the parafovea were labeled. Peroxidase injections into the superior colliculus produced labeling of scattered ganglion cells of all sizes in the retina, although no labeled cells were found within the centralmost 10 degrees eccentricity. From these observations it is concluded that all ganglion cells of the macaque retina project to the parvocellular layers of the dorsal lateral geniculate, but that only the largest ganglion cells of the more peripheral retina and not all cells of the parafovea project to the magnocellular laminae. In contrast, only scattered ganglion cells, although these are of all sizes, appear to project to the superior colliculus. Two major problems with the peroxidase tracing technique are described: 1. The extent of stainable peroxidase activity around the injection site appears to be larger than the area of injected tracer actually available for uptake by axons to produce labeled cells. 2. Cut or damaged axons appear to incorporate peroxidase sufficiently to produce labeling of the cell body.

Antidromic responses of orthodromically identified ganglion cells in monkey retina

1. Conduction velocities of two types of on-centre monkey ganglion cells, called phasic and tonic, have been measured by stimulating their axons in the optic tract while recording from their cell bodies in the retina.2. The average conduction velocity of twenty-two phasic and twenty-seven tonic cells is 3.8+/-S.D. 0.6 and 1.8+/-S.D. 0.4 m/sec respectively. Since the latter, but not the former, show opponent-colour responses, retinal signals carrying information about colour appear to be travelling in smaller axons than those not handling such information.3. Stimulation of the optic tract elicits several graded intraretinal potentials, which are negative in the optic nerve fibre layer and positive in the inner plexiform layer. One of these potentials, which is largest near the fovea, occurs simultaneously with the antidromic impulses of tonic ganglion cells and is considered to result from extracellular current generated by these cells.4. Stimulation of the optic tract suppresses the orthodromic responses of ganglion cells, more for phasic than for tonic ones. This suppression is only observed after a cell is antidromically driven and is considered most likely due to a transient hyperpolarization of the cell's membrane potential following an impulse.