The neuroanatomy of amnesia: amygdala-hippocampus versus temporal stem

Structural brain correlates of anterograde memory deficits in multiple sclerosis

Progressive decline of anterograde memory functions has been increasingly recognized as a frequent symptom in chronic multiple sclerosis. In order to investigate the brain structures involved, magnetic resonance imaging was performed in 20 patients. Neuropsychological assessment included the WAIS and WMS subtests information, picture completion, similarities, digit span, logical memory, and paired associate learning. All patients with severely impaired memory functions (n = 5) showed bilateral lesions in the medial temporal lobe, whereas in those patients with moderate (n = 10) or no measurable impairment of memory testing (n = 5) either no lesions were seen in the medial temporal lobes or these lesions were restricted to one side. A post hoc cluster analysis strikingly confirmed these results. The differences could not be related to the age of the patients, the disease duration, or the level of education. Extensive lesions in the white matter of the frontal lobes, thinning and lining of the corpus callosum, and bilateral involvement of the anterior cingulate gyrus had no bearing on the neuropsychological results. These findings indicate that bilateral demyelination in the hippocampal regions is the most likely explanation for the impairment of anterograde memory in such patients.

Preserved crossmodal association following bilateral amygdalotomy in man

The present study investigated the role of the amygdala in crossmodal association in man. Recent experimental evidence indicates amygdalectomized monkeys show deficits when required to identify, through vision, objects that have been previously examined by touch. We employed a delayed nonmatching to sample task pre- and postoperatively in a patient who underwent bilateral stereotaxic amygdalotomy for intractable aggressivity. Bilateral amygdalotomy did not impair our patient's ability to associate an object from one sensory modality to another, following a short delay. The results suggest that sensory association memory does not depend upon the amygdala in humans. Examination of the neuroanatomic structures affected by the differing surgical procedures in the two species revealed entorhinal and perirhinal cortices were damaged in amygdalectomized monkeys but not in the present case. Because entorhinal cortex is an important structure for the formation of new memory traces, it is possible that damage to the entorhinal and surrounding cortical regions, rather than damage to the amygdala itself, is responsible for the crossmodal association deficit in monkeys.

Age-related reduction in responses of rat hippocampal neurons to locally applied monoamines

Age-related changes in the responsiveness of hippocampal pyramidal neurons to norepinephrine (NE) and serotonin (5HT) were investigated using electrophysiological techniques. Local application of each monamine via pressure micro-ejection was employed to establish the dose which elicited a 50% change in spontaneous discharge rate of single pyramidal neurons; these data were used to construct dose response curves for the population of neurons tested in 3-6, 11-13, 18-20, and 27-30 month old rats. The percentage of cells responding in rats 18-20 and 27-30 months old decreased for both NE and 5HT. There was also a progressive increase with age in the population ED50 for 5HT starting at 18-20 months. For neurons which demonstrated a response to NE, no decrease in the population ED50 was observed. Taken together these data indicate that there is a progressive age-related decline in the postsynaptic response to NE and 5HT in the rodent hippocampus.

Dissociation of the effects of inferior temporal and limbic lesions on object discrimination learning with 24-h intertrial intervals

Monkeys with bilateral ablations of the inferior temporal cortical area TE were trained on a visual discrimination task thought to measure non-cognitive habit formation. The task consisted of 20 object discriminations presented concurrently, but at the rate of only one trial on each per day; successive trials on a given discrimination were thus separated by 24-h intertrial intervals. Performance on this task by the animals with TE lesions was compared to that of both normal control monkeys and monkeys that had sustained bilateral removals of the amygdala and hippocampus. In contrast to the latter animals, which learned the 24-h intertrial interval task about as quickly as the normal controls, monkeys with area TE removals were markedly impaired. Taken together with earlier findings demonstrating that ablation of area TE impairs visual recognition memory, the present results suggest that area TE contributes not only, like limbic structures, to a cognitive memory system, but also, unlike limbic structures, to a non-cognitive habit system. Evidence is reviewed suggesting that this latter system may involve a corticostriatal circuit.

The performance of amnesic subjects on tests of experimental amnesia in animals: delayed matching-to-sample and concurrent learning

A group of amnesic Korsakoff subjects and a group of alcoholic controls were trained on a test of visual recognition, delayed matching-to-sample with trial unique stimuli. This test was modelled on comparable tasks used in the development of animal models of human amnesia. It was found that the Korsakoff subjects were severely impaired when the task difficulty was increased by lengthening the retention delay beyond 10 sec or by increasing the number of items intervening between sample presentation and test. The amnesic subjects were also impaired on the acquisition of a set of concurrent visual discriminations. These results bear clear similarities to those obtained from experimental amnesic syndromes in monkeys.

The entorhinal cortex of the monkey: II. Cortical afferents

The entorhinal cortex of the monkey is commonly viewed as the major link between the cerebral cortex and the other fields of the hippocampal formation. Until recently, however, little was known about the origins of the cortical projections to the entorhinal cortex, and most of the available information is still based on degeneration studies. We have carried out a systematic analysis of these connections by placing small injections of the retrograde tracer wheat germ agglutinin conjugated to horseradish peroxidase into each of the fields of the entorhinal cortex of the Macaca fascicularis monkey. Retrogradely labeled cells were observed in several areas of the frontal and temporal lobes, the insula, and the cingulate cortex. In the frontal lobe, the greatest number of labeled cells were observed in the orbital region and specifically in areas 13 and 13a: labeled cells were also seen in areas 14, 11, and 12. In the dorsolateral frontal cortex, labeled cells were observed mainly in the rostral half of area 46; occasionally cells were also seen in areas 9, 8, and 6. In the cingulate cortex, labeled cells were observed in area 25, area 32, and rostral levels of area 24; fewer cells were observed at caudal levels of area 24 or in area 23. The retrosplenial region (areas 30 and 29), including its caudal extension along the rostral calcarine sulcus and its ventral extension into the temporal lobe, contained numerous labeled cells. In the temporal lobe, retrogradely labeled cells were arranged in two rostrocaudally oriented bands. Rostral to the hippocampal formation, the first band encompassed the piriform and periamygdaloid cortices and areas 35 and 36; the labeling in area 36 was continuous to the temporal pole. At more caudal levels this band was located immediately lateral to the hippocampal formation and included areas 35 and 36 rostrally and areas TH and TF caudally. The second band was situated in the superior temporal gyrus where labeled cells were observed in several distinct cytoarchitectonic fields, including the parainsular cortex in the fundus of the inferior limiting sulcus. In the insula proper, retrogradely labeled cells were seen mainly in the rostral or agranular division; far fewer were observed in the dysgranular and granular insula. Whereas there is little available physiological information concerning many of the cortical regions that project to the entorhinal cortex, on anatomical grounds they may be generally characterized as polysensory associational regions.

Separation of hippocampal and amygdaloid involvement in temporal memory dysfunctions

The role of the hippocampus and the amygdala in timing and in the memory of previously timed events was investigated in rats. Two testing procedures used the peak time (the time at which the maximum response rate occurred) to identify the time at which the rat expected reinforcement. Amygdala (AMG) lesions had no effect on the remembered time of reinforcement or on the ability to remember the duration of a previous stimulus. Fimbria-fornix (FF) lesions had two effects: these rats remembered the time of reinforcement as occurring earlier than it really did, and could not remember the duration of a previous stimulus even after a gap of only 0.5 s. This behavior pattern endured throughout testing in spite of reinforcement contingencies designed to eliminate it. Atropine, 0.45 mg/kg, caused control rats to forget the duration of a previous stimulus, while haloperidol, 0.15 mg/kg, did not. Taken together, these data indicate that the hippocampus, but not the amygdala, has an important role in the memory for time. They suggest that alterations in temporal processes may be intimately involved in the amnesic syndrome seen following damage to temporal lobe structures.

The performance of visual tasks while segments of the inferotemporal cortex are suppressed by cold

Cold was used to suppress the function of subdivisions of the inferotemporal cortex. Three cryodes were placed bilaterally, one over the lower bank of the superior temporal sulcus (sts), one over the middle temporal gyrus (mtg) and one over the inferior temporal gyrus (itg). The animals were tested with delayed match-to-sample (DMS) and simultaneous visual discriminations. The DMS required the animal to recall a projected image of an object over delays of 0, 15, 30 and 45 s. The 3 cryodes were cooled separately during the performance of the DMS and only itg cooling produced a deficit. This was compared to the effects of ablative bilateral lesions; damage to itg but not mtg disrupted performance of DMS. The greatest deficit was in an animal with a small lesion in the ventral pole and anterior extreme of itg. Cooling individual cryodes was without effect on a discrimination between horizontal and vertical stripes, but produced a significant deficit from each of the 3 placements on a discrimination between monkey faces. Chance performance on all visual discriminations resulted from cooling all cryodes. Unilateral cooling of all cryodes produced significant effects on the face discrimination, but there was no significant difference between the two sides in the severity of the deficit.

Vasopressin in Alzheimer's disease: a study of postmortem brain concentrations

Vasopressin (AVP) and its analogues are reported to improve learning- and memory-related performance in experimental animals, and perhaps also in humans. Memory impairment is a clinical hallmark of the dementing disorder, Alzheimer's disease. We have examined AVP concentrations in postmortem brain tissue from 12 patients with histologically confirmed Alzheimer's disease and 13 control subjects. AVP was measured by a highly specific and sensitive radioimmunoassay, validated by parallel inhibition curves and high-performance liquid chromatography. Alzheimer brains had either normal or slightly increased AVP levels in the neocortex, which does not have AVP cell bodies. Significant reductions in AVP content were found in the hippocampus, nucleus accumbens, and globus pallidus interna. Levels were normal in all other regions studied. Abnormalities of the brain vasopressin system may contribute to the memory deficit associated with Alzheimer's disease.

Visual learning and retention examined with reversible cold lesions of the anterior temporal lobe

Learning and retention of visual discriminations and delayed match-to-sample (DMS) performance were examined in monkeys while cooling the anterior temporal lobe. Four cryodes were bilaterally implanted on the dura overlying the anterior temporal cortex, an anterior pair covered the temporal pole (TP) and a posterior pair covered the anterior inferior temporal cortex (AIT). The visual discriminations were examined under 4 different test combinations of cooling and not cooling the anterior temporal lobe. Learning deficits were produced by cooling either TP or AIT. Once learned, there was no difficulty recalling discriminations under cooling or control conditions for either TP or AIT. There was a deficit during cooling in the recall of discriminations that had been learned prior to cooling TP or AIT. The animals were then trained and tested on a DMS task at a 0-s and 10-s delay. They performed at chance when either TP or AIT was cooled in the 0-s delay. Only TP was cooled at the 10-s delay and it also resulted in chance performance. The cold lesions demonstrated that the anterior temporal cortex, i.e. TP and AIT, has an important role in the processes of learning and, to a lesser extent, retention of visual information. The results also support previous findings regarding the participation of this area in DMS performance. The findings were discussed in relation to the amnesic syndrome.

Topography of projections from the medial prefrontal cortex to the amygdala in the rat

The projections from the rat medial prefrontal cortex to the amygdaloid complex were investigated using retrograde transport of fluorescent dyes and anterograde transport of horseradish peroxidase-WGA. The ventral anterior cingulate, prelimbic, infralimbic and medial orbital areas and the taenia tecta were found to project to the amygdaloid complex. The projections from the prelimbic area arose bilaterally. The medial orbital, prelimbic and anterior cingulate areas send convergent projections to the basolateral nucleus. The prelimbic area has additional projections to the posterolateral cortical nucleus and amygdalo-hippocampal area. The infralimbic area does not project to the basolateral nucleus and cortico-amygdaloid projections from this area are focussed on the anterior cortical nucleus and the anterior amygdaloid area. Both prelimbic and infralimbic areas project to an area situated between the central, medial and basomedial nuclei. Based on similar projections, this area appears to be a caudal continuation of the anterior amygdaloid area. The results indicate that the medial prefrontal component of the "basolateral limbic circuit" is restricted to the anterior cingulate and prelimbic areas. No evidence was obtained to support the existence of a medial prefronto-amygdaloid component of the "visceral forebrain".

Mechanisms of memory

Recent studies of animals with complex nervous systems, including humans and other primates, have improved our understanding of how the brain accomplishes learning and memory. Major themes of recent work include the locus of memory storage, the taxonomy of memory, the distinction between declarative and procedural knowledge, and the question of how memory changes with time, that is, the concepts of forgetting and consolidation. An important recent advance is the development of an animal model of human amnesia in the monkey. The animal model, together with newly available neuropathological information from a well-studied human patient, has permitted the identification of brain structures and connections involved in memory functions.

Modulating effects of posttraining epinephrine on memory: involvement of the amygdala noradrenergic system

These experiments examined the effects, on retention, of posttraining intra-amygdala administration of norepinephrine (NE), and propranolol. Rats were trained on a one-trial step-through inhibitory avoidance task and tested for retention 24 h later. Injections were administered bilaterally (1.0 microliter/injection) through chronically-implanted cannulae. Low doses of NE (0.1 or 0.3 microgram) administered shortly after training enhanced retention while higher doses (1.0 or 5.0 micrograms) were ineffective. Retention was not affected by NE administered 3 h after training. The effect of intra-amygdala NE on retention is blocked by simultaneous administration of propranolol (0.2 microgram). This finding suggests that the memory-enhancing effect of NE may be mediated by beta-receptors. Posttraining intra-amygdala NE also attenuated the retention deficit produced by adrenal demedullation. Further, intra-amygdala injections of propranolol (0.2 microgram) blocked the enhancing effect, on retention, of posttraining s.c. injections of epinephrine. These findings suggest that activation of noradrenergic receptors in the amygdala may be involved in memory processing and may play a role in the memory-modulating effect of peripheral epinephrine.

Transient global amnesia: a case-control study

The clinical findings and laboratory results from 18 patients with clear-cut transient global amnesia (TGA) were compared with the results from 90 nonischemic neurological control patients using case-control analysis. A prior episode of cerebral ischemia, generally in the posterior circulation, was the most significant risk factor for TGA. Other risk factors for cerebrovascular disease were more prominent in the TGA group, but this association became less significant when cerebral ischemia was controlled for. Migraine, epilepsy, and psychiatric disorders were not significantly associated with TGA. In 5 of 13 patients, computed tomography showed focal thalamic and temporal lobe abnormalities. The recurrence rate was 7.0% for both TGA and subsequent cerebral ischemia. We conclude that TGA is closely linked to cerebrovascular disease. Further, prior damage to anatomical structures critical to memory may be necessary for the expression of this syndrome.

Clinical history, brain metabolism, and neuropsychological function in Alzheimer's disease

Data concerning 7 patients with a diagnosis of presumptive Alzheimer's disease (mean age, 65.6 years) are presented in detail in relation to the patients' regional cerebral metabolic rates for glucose. Rates were measured by positron emission tomography with fluorine 18-labeled fluoro-2-deoxy-D-glucose under conditions of reduced visual and auditory stimulation. A relationship was found between severity of dementia and brain metabolism. In patients with mild to moderate Alzheimer's disease, memory and intellectual deficits were evident without major reductions in absolute metabolic rates, while ratios of regional to whole brain metabolism revealed reductions in regions of the parietal lobes. In the late, severe form of the disease, brain metabolic rates were consistently and significantly reduced. The findings suggest that memory and intellectual deficits are reflected in reductions of brain metabolism in some brain regions in mild to moderate forms of Alzheimer's disease and that, in the late, severe form of the disease, reductions occur consistently throughout the brain.

Bilateral and limited amygdalohippocampal lesions causing a pure amnesic syndrome

A pure amnesic syndrome of 21 months' duration occurred in a 36-year-old man following a transient confusional state. The patient died of Hodgkin's disease. At postmortem examination, bilateral and symmetrical neuronal loss, without inflammatory changes, was restricted to the hippocampus and amygdaloid bodies.

Hypotheses on mnemonic information processing by the brain

From a neuropsychological point of view, hypotheses are offered on the possible action of the brain in the processing of mnemonic information for long-term storage (or for retrieval of long-term stored information). It is argued that strict relations between damage of circumscribed brain structures and amnesia, as they have been suggested in recent case reports, are questionable for several reasons: Firstly, the involved regions differ between cases; secondly, there is some counter-evidence from other cases in which similar neuronal damage failed to result in lasting amnesic disturbances; thirdly, it is hypothesized that even from circumscribed brain damage it is not justifiable to conclude that the lesioned structure is solely or principally responsible for the observed mnemonic changes, as the brain acts in an integrative way, that is, on the basis of a wide-spread network of neuronal information processing. On the basis of these and related arguments, hypotheses and models on mnemonic information processing in the intact and in the damaged brain are derived. With these hypotheses even the frequent observation of interindividual differences in mnemonic information processing finds a possible explanation which is in conformity to known anatomical circuits and connections and to principles of neuronal coding.

Amnesia in monkeys after lesions of the mediodorsal nucleus of the thalamus

Recent successes in developing an animal model of human amnesia in the monkey have made it feasible to try to identify with certainty the specific structures in the diencephalon and medial temporal region that cause amnesia when damaged. Monkeys with small lesions restricted largely to the posterior portion of the mediodorsal nucleus of the thalamus were given a test of memory sensitive to human amnesia and a second test that is analogous to the skill-based tasks performed normally by amnesic patients. The monkeys exhibited a marked impairment on the first test and performed normally on the second. The results show that circumscribed lesions of the mediodorsal nucleus can cause substantial amnesia.

Hippocampal unit activity and delayed response in the monkey

Single unit activity was recorded from the hippocampus while Japanese monkeys (Macaca fuscata, n = 4) were performing a delayed response (DR) task. A total of 272 units showed an obvious change in discharge rate in relation to the events of the DR task. These 272 related units were classified into 6 groups: cue-light related units (n = 24), cue- and choice-light related units (n = 41), choice-light related units (n = 21), response-related units (n = 51), reward-error units (n = 17), and delay units (n = 118). Reward-error units contained reward-related and error-related units. Error-related units showed changes in firing after incorrect responses and/or after omission of reward on correct trials. It is noteworthy that 43.4% of the related units are delay units which showed increased or decreased firing preferentially during the delay period. Some units showed a differential firing pattern during cue or delay period depending on the spatial position of the cue. The results of the present study are interpreted as an experimental evidence for the involvement of the hippocampus in DR task.

The neuropathology of amnesia

Relations between brain damage and memory disturbance are outlined with emphasis on the so-called amnesic syndrome. Following a brief introduction into forms of memory and memory failures, the basic causes of brain damaage (with relevance to amnestic failures) are described. Thereafter, the two best-known forms of brain damage-amnesia relations are reviewed: the consequences of damage to medial temporal lobe structures and to diencephalic regions. For the cases with medial temporal lobe damage, evidence is reported in greater detail for H.M., who has been examined more than any other amnesic patient for more than 30 years now, as a considerable amount of literature has accumulated on his behavior in diverse situations. Other cases with more or less circumscribed damage to medial temporal lobe structures are reviewed so as to outline criteria for or against the hypothesis that there are regions within the medial temporal lobe whose damage might be critical for the amnesic syndrome. Two cases of diencephalic amnesia are summarized in particular (cases of Mair et al., 1979) as they have received extensive neuropsychological and neuropathological investigation. Other cases with, for example, Korsakoff's disease are reviewed, as well as cases with diencephalic, or combined mesencephalic-diencephalic damage without nutritional causes. A third group of patients with massive, but still selective amnesic disturbances are then described: cases of basal forebrain damage, followed by descriptions of Alzheimer's disease which has similarities in the underlying neuropathology. This leads over to cases with more generalized intellectual deteriorations (dementia), which may have developed on the basis of primarily cortical damage or damage principally to basal ganglia structures. After reviewing cases with mainly material-specific memory failures--usually as a consequence of restricted neocortical damage--a separate section follows on patients in whom retrograde amnesia is the prominent symptom. The contribution of animal models of human amnesia is critically reviewed and discrepancies are analyzed between human and animal memory disturbances. This section emphasizes the value of investigating inter-dependencies between brain structures by pointing out that relations between memory disturbances and brain damage may be more complicated than apparent from a simple structure-function assignment. This aspect is further followed up in the conclusions.

Can amnesia be caused by damage of a single brain structure?

The relation of amnesia and damage to some particular brain regions is discussed by reviewing the main findings of selected human case reports. It is argued that frequently a too straightforward and unidimensional interpretation of the relations between brain damage and a behavioral deficit is formulated in such reports. Evidence obtained by modern anatomical techniques as well as the widespread and time-dependent effects of lesions make it necessary to consider a lesion of a particular structure of the brain and correlated mnemonic disturbances as possibly due to an altered equilibrium in an extensive network of the brain. The primary lesion of one or the other specific structure may lead to severe and lasting amnesia or may fail to do so depending on its influence on other brain regions.

Severe tactual memory deficits in monkeys after combined removal of the amygdala and hippocampus

Monkeys with bilateral removals of both the amygdaloid complex and hippocampal formation were far more severely impaired on a tactual memory task than were monkeys with removal of either structure alone. These data parallel earlier findings on visual memory in monkeys and suggest that: (i) the memory deficit following combined ablations of the amygdala and hippocampus is multimodal; and (ii) the global anterograde amnesia observed in patients with medial temporal-lobe damage is also due to combined damage to these two structures.

Convergence of basolateral amygdaloid and mediodorsal thalamic projections in different areas of the frontal cortex in the rat

The extent of convergence of mediodorsal thalamic and amygdalar afferents on the rat's frontal cortex was studied by tracing retrogradely labeled cells following injections of horseradish peroxidase (HRP). HRP was applied iontophoretically in extremely small injections throughout all areas of the frontal cortex. The following organization was revealed: Converging inputs from the mediodorsal nucleus and the amygdala are observed in the posterior parts of the pre- and infralimbic areas, in the posterior half of the dorsal and ventral agranular insular areas and in the lateral and dorsal precentral areas. Both mediodorsal and amygdaloid afferents reach the dorsal tip of the frontal cortex. Only the mediodorsal afferents were found to terminate in the anterior parts of the pre- and infralimbic areas and in the anterior part of the dorsal division of the anterior cingulate area and in the medial precentral area. On the lateral side of the hemisphere the anterior halves of the dorsal and ventral agranular insular areas receive mediodorsal afferents. Amygdaloid, but not mediodorsal afferents, were found following injections into the more posterior parts of the lateral precentral area. These results are discussed with respect to the extent of the prefrontal cortex in the rat and its definability as a target area of subcortical nuclei. Functional aspects of the anatomical convergence of connections within the so-called basolateral limbic circuit are outlined.

Visual recognition impairment following medial thalamic lesions in monkeys

Monkeys with surgical lesions which removed the medial portions of the medial and anterior thalamic nuclei were markedly impaired on a test of object recognition. The same animals were able to learn visual pattern discriminations and a spatial delayed response task at a normal rate. These findings indicate that lesions in the medial thalamus produce a selective impairment in visual recognition memory in monkeys and, consequently, may provide an experimental model for human "diencephalic amnesia".