Mediated generalization in discrimination learning by Rhesus monkeys.

Monkeys received discrimination training in which the choice of either a simultaneous or a serial compound of two visual images was rewarded before we assessed the monkeys' conditioned preference for one of the images from the compound. This preference was reduced or blocked if the other image had been associated with reward rather than nonreward prior to compound training. By contrast, the preference was enhanced if the other image was associated with reward rather than nonreward after compound training. The magnitudes of the blocking and enhancement were unaffected by the temporal structure, simultaneous or serial, of the stimulus compound. These results are discussed in terms of the representation of stimulus compounds, the role of within-compound associations, and the similarities between serial and simultaneous visual compounds as processed by monkeys.

Addition of fornix transection to frontal-temporal disconnection increases the impairment in object-in-place memory in macaque monkeys.

Both frontal-inferotemporal disconnection and fornix transection (Fx) in the monkey impair object-in-place scene learning, a model of human episodic memory. If the contribution of the fornix to scene learning is via interaction with or modulation of frontal-temporal interaction--that is, if they form a unitary system--then Fx should have no further effect when added to frontal-temporal disconnection. However, if the contribution of the fornix is to some extent distinct, then fornix lesions may produce an additional deficit in scene learning beyond that caused by frontal-temporal disconnection. To distinguish between these possibilities, we trained three male rhesus monkeys on the object-in-place scene-learning task. We tested their learning on the task following frontal-temporal disconnection, achieved by crossed unilateral aspiration of the frontal cortex in one hemisphere and the inferotemporal cortex in the other, and again following the addition of Fx. The monkeys were significantly impaired in scene learning following frontal-temporal disconnection, and furthermore showed a significant increase in this impairment following the addition of Fx, from 32.8% error to 40.5% error (chance = 50%). The increased impairment following the addition of Fx provides evidence that the fornix and frontal-inferotemporal interaction make distinct contributions to episodic memory.

Associative recognition in a patient with selective hippocampal lesions and relatively normal item recognition.

Previous work (Mayes et al., Hippocampus 12:325-340, 2002) found that patient YR, who suffered a selective bilateral lesion to the hippocampus in 1986, showed relatively preserved verbal and visual item recognition memory in the face of clearly impaired verbal and visual recall. In this study, we found that YR's Yes/No as well as forced-choice recognition of both intra-item associations and associations between items of the same kind was as well preserved as her item recognition memory. In contrast, YR was clearly impaired, and more so than she was on the above kinds of recognition, at recognition of associations between different kinds of information. Thus, her recognition memory for associations between objects and their locations, words and their temporal positions, abstract visual items or words and their temporal order, animal pictures and names of professions, faces and voices, faces and spoken names, words and definitions, and pictures and sounds, was clearly impaired. Several of the different information associative recognition tests at which YR was impaired could be compared with related item or inter-item association recognition tests of similar difficulty that she performed relatively normally around the same time. It is suggested that YR's familiarity memory for items, intra-item associations, and associations between items of the same kind was mediated by her intact medial temporal lobe cortices and was preserved, whereas her hippocampally mediated recall/recollection of these kinds of information was impaired. It is also suggested that the components of associations between different kinds of information are represented in distinct neocortical regions and that initially they only converge for memory processing within the hippocampus. No familiarity memory may exist in normal subjects for such associations, and, if so, YR's often chance recognition occurred because of her severe recall/recollection deficit. Conflicting data and views are discussed, and the way in which recall as well as item and associative recognition need to be systematically explored in patients with apparently selective hippocampal lesions, in order to resolve existing conflicts, is outlined.

The effect of cingulate cortex lesions on task switching and working memory.

Anatomic interconnections between the prefrontal and anterior cingulate cortices suggest that these areas may have similar functions. Here we report the effect of anterior cingulate removal on task switching, error monitoring, and working memory. Neuroimaging studies have implicated the cingulate cortex in all these processes. Six macaques were taught task switching (TS) and delayed alternation (DA) paradigms. TS required switching between two conditional response tasks with mutually incompatible response selection rules. DA required alternation between two identically covered food-well positions. In the first set of experiments, anterior cingulate lesions did not consistently impair TS or DA performance. One animal performed worst on both TS and DA and in this animal the cingulate sulcus lesion was most complete. In the second set of experiments, we confirmed that larger anterior cingulate lesions, which included the sulcus, consistently impaired TS but only led to a mild and equivocal impairment of DA. The TS error pattern, however, did not suggest an impairment of TS per se. The consequence of a cingulate lesion is, therefore, distinct to that of a prefrontal lesion. TS error distribution analyses provided some support for a cingulate role in monitoring responses for errors and subsequent correction but the pattern of reaction time change in TS was also indicative of a failure to sustain attention to the task and the responses being made.

The effect of cingulate lesions on social behaviour and emotion.

Functional and structural neuroimaging of the human cingulate cortex has identified this region with emotion and social cognition and suggested that cingulate pathology may be associated with emotional and social behavioural disturbances. The importance of the cingulate cortex for emotion and social behaviour, however, has not been clear from lesion studies. Bilateral lesions in the cingulate cortex were made in three macaques and their social interactions were compared with those of controls. Subsequently, cingulate lesions were made in the three controls and their behaviour was compared before and after surgery. Cingulate lesions were associated with decreases in social interactions, time spent in proximity with other individuals, and vocalisations but an increase in manipulation of an inanimate object. The results are consistent with a cingulate role in social behaviour and emotion.

The anterior cingulate and reward-guided selection of actions.

Macaques were taught a reward-conditional response selection task; they learned to associate each of two different actions to each of two different rewards and to select actions that were appropriate for particular rewards. They were also taught a visual discrimination learning task. Cingulate lesions significantly impaired selection of responses associated with different rewards but did not interfere with visual discrimination learning or performance. The results suggest that 1) the cingulate cortex is concerned with action reward associations and not limited to just detecting when actions lead to errors and 2) that the cingulate cortex's function is limited to action reinforcer associations and it is not concerned with stimulus reward associations.

The role of the vertical meridian in visual memory for objects.

It is widely believed that, in human and nonhuman primates, visual memories of objects are stored in the temporal lobe. Electrophysiological results in monkeys, however, indicate that when a visual scene contains two or more objects, with at least one object in each visual hemifield, neurons in the temporal lobe of each hemisphere respond only to the objects that are in the contralateral visual hemifield, and their activity is unaffected by the objects in the ipsilateral hemifield. Putting these two premises together predicts that object memory should fail, or at least suffer a substantial decrement, when an object is presented for learning and retention as part of such a scene, but crosses the vertical meridian between the learning trial and the retention test. The effect of this change should be much greater than the effect of an equal retinal translation that crosses the horizontal rather than the vertical meridian. An experiment with normal human subjects verified this prediction under conventional conditions of tachistoscopic viewing, with a single constant fixation spot. A further condition in the same experiment, however, tested the same retinal translations in a more naturalistic condition, where the retinal changes were produced by varying the position on the display screen of the fixation spot rather than of the objects. Here, there was no significant special effect of crossing the vertical meridian. We conclude that visual memories are not stored exclusively in the temporal lobe.

Unilateral lesions of the cholinergic basal forebrain and fornix in one hemisphere and inferior temporal cortex in the opposite hemisphere produce severe learning impairments in rhesus monkeys.

It has been proposed that isolation of the inferior temporal cortex and medial temporal lobe from their cholinergic afferents results in a severe anterograde amnesia. To test this hypothesis directly, seven rhesus monkeys received a unilateral immunotoxic lesion of the cholinergic cells of the basal forebrain with an ipsilesional section of the fornix. In a second surgery, inferior temporal cortex was ablated in the opposite hemisphere. All animals were severely impaired at learning visual scenes and object-reward associations. The impairment in learning scenes was correlated with cholinergic cell loss in the basal forebrain, but not with generalized tissue damage. Two monkeys served as surgical controls with saline injection in place of the immunotoxin, but all other procedures the same, and were not as severely impaired as those with immunotoxic lesions. Previous work has shown that monkeys with bilateral section of the anterior temporal stem (white matter of the temporal lobe), amygdala and fornix show a severe new learning impairment, and provide a model of human medial temporal lobe amnesia. One effect of this combined ablation is to isolate inferior temporal cortex and medial temporal lobe from their cholinergic afferents, possibly in addition to a direct disruption of the hippocampal system. The results of the present study, then, provide a novel link between the mechanisms of medial temporal lobe amnesia and Alzheimer's disease in which the cholinergic basal forebrain shows pathology. We propose that in both cases the mnemonic impairments result from isolating inferior temporal cortex and medial temporal lobe from their cholinergic afferents, possibly in addition to a direct disruption of the hippocampal system.

Selective perceptual impairments after perirhinal cortex ablation.

It has been suggested that the primate perirhinal cortex contributes exclusively to memory. However, recent studies in macaque monkeys have implied that the perirhinal cortex may also contribute to object perception. To investigate whether the perirhinal cortex does contribute to perception, we devised several perceptual oddity tasks in which monkeys had to choose which stimulus of several presented concurrently on a touch screen was different. Macaques with bilateral perirhinal cortex ablations were selectively impaired relative to controls at perceptually discriminating the odd stimulus when the odd stimulus was a different object and when the discrimination could not be done on the basis of simple differences in features between the stimuli. They remained unimpaired relative to controls on discriminating the odd stimulus when the odd stimulus was a different color, a different shape, or a different size even when these discriminations were extremely difficult. They were also impaired on human and monkey face oddity tasks and oddity tasks with scenes containing objects. Therefore, we reject the notion that the macaque perirhinal cortex has a role exclusive to memory and conclude that the macaque perirhinal cortex does contribute to perception. We argue that the perirhinal cortex is neither specialized for perception nor memory processes alone, but rather, is specialized for processing stimuli that require processing at a more abstract level such as at the level of an object and that the perirhinal cortex contributes to both memory and perception of such stimuli.

What is a memory system? Horel's critique revisited.

J.A. Horel's critique of what he termed "the hippocampal memory hypothesis" turns out, 23 years later, to have been remarkably discerning and prophetic. There is now an overwhelming weight of evidence to confirm his four key proposals: that selective destruction of the hippocampus or fornix does not produce dense global amnesia; that the effects of hippocampal or fornix lesions are not primarily a memory impairment, but an impairment in processing spatial information; that damage to the anterior temporal stem is part of the explanation of dense temporal lobe amnesia; and that the interaction of temporal cortex with prefrontal cortex is essential in memory. This review summarizes the modern evidence that reinforces each of these four proposals. A final section argues that, not only in the case of the hippocampus but also in the case of other temporal and frontal cortical areas that are involved in normal memory, the concept of a "memory system" is harmful.

Visual agnosia and Klüver-Bucy syndrome in marmosets (Callithrix jacchus) following ablation of inferotemporal cortex, with additional mnemonic effects of immunotoxic lesions of cholinergic projections to medial temporal areas.

Inferotemporal ablations in the New World monkey, the common marmoset (Callithrix jacchus), produced a persistent impairment on visual discrimination learning and a florid, but transient, Klüver-Bucy syndrome. Monkeys with these ablations were impaired on acquisition of object discriminations to a high criterion and on concurrent discrimination learning, to a single high criterion across all trials. Neither the control monkeys nor the monkeys with inferotemporal ablations found acquisition more difficult when the component discriminations of a set were presented concurrently compared to consecutively, although the monkeys with inferotemporal ablations found acquisition under both these conditions somewhat more difficult than did control monkeys. This suggests that the severe impairment caused by inferotemporal ablations on concurrent learning measured across all trials is due to the need for sustained performance across a concurrent set rather than to the extra mnemonic demands of concurrent presentation. When immunotoxic lesions of the cholinergic projection to the hippocampal formation were added to the inferotemporal ablations, a further impairment on retention, and a differential impairment on concurrent, compared to consecutive, learning was observed. Previous studies have shown that lesions of the cholinergic projection to the hippocampus alone, or excitotoxic hippocampal lesions, do not affect simple visual discrimination learning. It is suggested that large inferotemporal ablations in monkeys produce a visual agnosia which causes severe 'psychic blindness' in the first instance, and a persistent impairment on visual discrimination learning. The hippocampus makes a contribution, which may be mnemonic, to discrimination performance after inferotemporal ablations.

Crossed unilateral lesions of medial forebrain bundle and either inferior temporal or frontal cortex impair object recognition memory in Rhesus monkeys.

In monkeys, section of the fornix, amygdala and anterior temporal stem results in a severe anterograde amnesia. Immunolesions of the cholinergic cells of the basal forebrain suggest that this amnesia is a result of isolating the inferior temporal cortex and medial temporal lobe from their cholinergic basal forebrain afferents. In this experiment, six monkeys were trained in a delayed match-to-sample task and then received a section of the medial forebrain bundle in one hemisphere and an ablation of either the frontal or inferior temporal cortex in the opposite hemisphere. All the animals were severely impaired in the performance of this task following this surgery, and the severity of the impairment was independent of the cortical area from which the medial forebrain bundle was disconnected. These results support a model of fronto-temporal interaction via the basal forebrain in new learning, in which midbrain sites related to reward modulate the cholinergic basal forebrain activity.

Visual discrimination learning impairments produced by combined transections of the anterior temporal stem, amygdala and fornix in marmoset monkeys.

Marmoset monkeys (Callithrix jacchus) with bilateral transections of the anterior temporal stem, amygdala and fornix were unable to relearn a 2-choice object discrimination first learnt prior to surgery, and were very severely impaired at relearning a concurrent object discrimination task which they had learnt and relearnt prior to surgery, indicating that they had a dense retrograde amnesia. They also had difficulty learning new visual object discriminations but were only mildly impaired on spatial learning. When tested on new learning of concurrent discriminations 8 to 10 weeks after surgery, three operated monkeys were unable to reach criterion in 400 trials while the remaining two operated monkeys performed within the normal range. The operated monkeys were subsequently shown to be impaired on acquisition of shape discriminations using black objects. These anterograde effects suggest that the impairment runs mainly in the domain of visual analysis. The monkeys also exhibited many of the features of the Klüver-Bucy syndrome. Histological analysis indicated that in addition to cutting some of the subcortical temporal lobe efferent pathways, the surgical procedures had cut the cholinergic afferents to the temporal neocortex, entorhinal cortex, and hippocampus. In a second experiment we found that treatment with the cholinergic agonist pilocarpine, which is effective in monkeys with specific cholinergic lesions, was unable to remediate the lesion-induced impairments. This suggests that transection of the non-cholinergic afferents, or the temporal lobe subcortical efferents, contributed to the behavioural syndrome and the learning and retention deficits seen in these monkeys.

Crossed unilateral lesions of the medial forebrain bundle and either inferior temporal or frontal cortex impair object-reward association learning in Rhesus monkeys.

In an accompanying paper we showed that combined transection of the fornix, amygdala and temporal stem in monkeys produced dense amnesia, including an impairment in visual object-reward association learning. We proposed that this combined surgical section had its effect by isolating temporal cortex from the ascending projections of the basal forebrain and midbrain structures. To test this hypothesis, in the present experiment we disconnected the inferior temporal cortex from these basal forebrain and midbrain structures, while sparing cortical white matter, by crossed unilateral lesions of the medial forebrain bundle in one hemisphere and inferior temporal cortex in the opposite hemisphere. The aim of the medial forebrain bundle lesion was to section axons of cells, both those that project to the cortex via the medial forebrain bundle, and those which control the activity of these same structures. A single unilateral lesion alone had no effect on the ability to learn and remember visual object-reward associations, but the crossed unilateral lesions produced an impairment in this task which was equal in severity to the impairment seen earlier after bilateral section of the fornix, amygdala and temporal stem. The impairment was not an effect of interrupting fibres to the cortex from the ventromedial hypothalamus, or of unilateral sensory neglect. This supports the hypothesis that these midbrain and basal forebrain afferents to the inferior temporal cortex are important for new visual learning. Furthermore, an impairment of equal severity was demonstrated in a separate group of animals that received crossed unilateral lesions of the medial forebrain bundle in one hemisphere and of the frontal cortex in the opposite hemisphere. We propose that the frontal cortex acts to modulate basal forebrain activity which in turn reinforces object representations in the inferior temporal cortex during learning.

Dense amnesia in the monkey after transection of fornix, amygdala and anterior temporal stem.

The traditional explanation of dense amnesia after medial temporal lesions is that the amnesia is caused by damage to the hippocampus and related structures. An alternative view is that dense amnesia after medial temporal lesions is caused by the interruption of afferents to the temporal cortex from the basal forebrain. These afferents travel to the temporal cortex through three pathways, namely the anterior temporal stem, the amygdala and the fornix-fimbria, and all these three pathways are damaged in dense medial temporal amnesia. In four experiments using different memory tasks, we tested the effects on memory of sectioning some or all of these three pathways in macaque monkeys. In a test of scene-specific memory for objects, which is analogous in some ways to human episodic memory, section of fornix alone, or section of amygdala and anterior temporal stem sparing the fornix, each produced a significant but mild impairment. When fornix section was added to the section of anterior temporal stem and amygdala in this task, however, a very severe impairment resulted. In an object recognition memory task (delayed matching-to-sample) a severe impairment was seen after section of anterior temporal stem and amygdala alone, with or without the addition of fornix section; this impairment was significantly more severe than that which was seen in the same task after amygdalectomy leaving the temporal stem intact, with or without fornix section. Animals with combined section of anterior temporal stem, amygdala and fornix were also impaired in object-reward association learning. However, the retention of pre-operatively acquired object-reward associations was at a high level. These results show that the pattern of impairments after section of anterior temporal stem, amygdala and fornix in the monkey, leaving hippocampus intact, resembles human dense amnesia and is different from the effects of hippocampal lesions in the monkey.

Perceptual and mnemonic matching-to-sample in humans: contributions of the hippocampus, perirhinal and other medial temporal lobe cortices.

Two questions were addressed by the present study. The first was whether the previously reported item recognition deficit which is shown by amnesic patients may be due to a perceptual rather than a memory deficit. To address this question a group of amnesic patients were tested on a 14-choice forced-choice visual item recognition test which included a "simultaneous" condition in which the sample remained visible during the matching decision and a zero second delay. Eacott, Gaffan and Murray (1994) have reported an impairment in simultaneous matching-to-sample following perirhinal damage in monkeys. In our amnesic patients, a deficit was found only after filled delays of 10 seconds or longer and this was also the case for a subgroup of patients whose damage included the perirhinal cortex. The second question, which arose from the model of Aggleton and Brown (1999), was whether performance on the DMS task would remain intact following selective damage to the hippocampus. We tested a patient with bilateral damage to the hippocampus on the 14-choice DMS task and found that her performance was not significantly impaired at delays of up to 30 seconds.

Mediodorsal thalamic function in scene memory in rhesus monkeys.

Three monkeys were trained preoperatively in a scene memory task which is analogous, in some ways, to human episodic memory. The same animals were also trained in object-reward association memory. Following bilateral ablations of almost the entire magnocellular division of the mediodorsal thalamic nucleus, the animals were impaired both in scene memory and in object-reward association memory. These results, combined with recent results in object recognition memory from monkeys with mediodorsal thalamic lesions, show that the impairment produced by this lesion is more general, affecting a broader range of memory tasks, than the impairment which is produced in monkeys by lesions restricted to the hippocampus-fornix-mamillary system. It is also more severe than the effect of lesions limited to the medial part of the magnocellular division of the mediodorsal thalamic nucleus. These findings extend the evidence that the magnocellular division of the mediodorsal thalamic nucleus has an important and general role in memory, and they are consistent with the proposal that lesions of the magnocellular division of that nucleus have their effect by disrupting the function of prefrontal cortex.

Differential cognitive effects of colloid cysts in the third ventricle that spare or compromise the fornix.

A series of twelve cases, all of whom had received surgery for the removal of a colloid cyst in the third ventricle, was examined on a series of memory tests. The only consistent predictor of poor memory performance that could be detected from MRIs was the presence of bilateral interruption of the fornix, which occurred in three of the subjects. Although these three cases were poor on tests of learning and recall, there was evidence that recognition was less impaired. The subjects were also tested on a set of recognition and concurrent discriminations that closely matched tests given to non-human primates. Clear parallels were found between the apparent effects of fornix damage in these clinical cases and those observed following more selective surgery in non-human primates. These findings not only indicate that fornix damage is sufficient to induce anterograde amnesia but also support the validity of recent animal tests that are thought to capture aspects of episodic memory.

Comparison of perirhinal cortex ablation and crossed unilateral lesions of the medial forebrain bundle from the inferior temporal cortex in the rhesus monkey: effects on learning and retrieval.

Seven monkeys learned new object-reward associations and scene problems and were overtrained on 100 problems of each type. Four monkeys received crossed lesions of the medial forebrain bundle (MFB) and inferior temporal cortex, with the later addition of a fornix section ipsilateral to the MFB lesion. The remaining 3 monkeys received bilateral perirhinal cortex ablation. Disconnection of the MFB from the inferior temporal cortex impaired postoperative new learning, but the retrieval of problems overtrained preoperatively was relatively preserved. Subjects with perirhinal cortex ablation were severely impaired in new learning and at the retrieval of scene problems, but retention of object-reward associations was relatively well preserved. The results support the hypothesis that isolation of the inferior temporal cortex from basal forebrain and midbrain afferents results in dense anterograde amnesia, whereas the role of the perirhinal cortex in learning is dependent on the perceptual difficulty of the task.

Idiothetic input into object-place configuration as the contribution to memory of the monkey and human hippocampus: a review.

Memory for object-place configurations appears to be a common function of the hippocampus in the human and monkey brain. The nature of the spatial information which enters into these object-configural memories in the primate, and the location of the memories themselves, have remained obscure, however. In the rat, much evidence indicates that the hippocampus processes idiothetic spatial information, an estimate of the animal's current environmental location derived from path integration. I propose that in primates the hippocampus provides idiothetic information about the environmental location of body parts, and that the main function of this information in the primate brain is to become configured with object-identity information provided by temporal lobe cortex outside the hippocampus.