Damage to diencephalic structures is stated to give rise to memory dysfunction. Amnesia is likely to occur following vascular lesions in the ventral portion of the thalamus. The CT findings of 6 of our own cases and 5 patients reported in the literature, all with selective vascular lesions of the thalamus, were studied to determine the critical structures involved in human memory processes more closely. Thalamic amnesia probably depends on intrathalamic white matter lesions more than on nuclear lesions. The mamillothalamic tract and the ventral portion of the lamina medullaris interna are the most likely candidates in the mediation of memory processes and a combined lesion of these structures may be responsible for thalamic amnesia in man. Two patients without significant memory dysfunction had lesions in the ventrobasal portion of the mediodorsal nucleus sparing the mamillothalamic tract and the ventral part of the lamina medullaris interna. Our findings correspond well with the understanding of amnesia as a 'disconnection syndrome' stressed recently by Warrington and Weiskrantz (1982) and with Mishkin's (1982) model of the memory system in monkeys.

We present nine experiments, in three study phases, which test the hypothesis that learning methods which prevent the making of errors ("errorless learning") will lead to greater learning than "trial-and-error " learning methods amongst individuals who are memory impaired as a result of acquired brain injury. Results suggest that tasks and situations which facilitate retrieval of implicit memory for the learned material (such as learning names with a first letter cue) will benefit from errorless learning methods, whilst those that require the explicit recall of novel associations (such as learning routes or programming an electronic Requests for reprints should be sent to Jonathan J. Evans,

The case of a patient with above-average intelligence and educational background, high motivation, and an approximate IQ-MQ difference of 40 points is documented. The patient has been examined repeatedly for nearly a decade. Extensive neuroradiological material of his focal bilateral brain damage in the dorsal diencephalon is available. A widespread range of cognitive tests was used to investigate his actual performance on all relevant aspects of intelligence, attention, subjective memory, immediate retention, learning, skill and problem solving abilities, concept formation, cognitive flexibility, priming, constructional ability, retrograde memory, and long-term retention. The total of more than 50 tests included German-language forms of the revised Wechsler Memory Scale and of the Rivermead Behavioural Memory Test. The patient's short-term memory and attention were, in spite of his advanced years, average or well above average. He gave a number of examples of still intact skills and implicit memory abilities, though there was no uniformity in his performance on implicit memory tests (e.g., with respect to stored vs. new implicit information). He had no awareness of his severe anterograde and retrograde amnesia, documented over a large range of verbal and figural tests. Taken together, the results from our patient confirm the principal dichotomy between declarative and nondeclarative mnestic functions, but give evidence for some restrictions as well. They furthermore demonstrate that focal diencephalic damage may result in profound anterograde and selective retrograde amnesia, especially with respect to data-based material, and that disconnecting portions of the medial and basolateral limbic circuits has devastating consequences on memory.

No abstract

Unilateral posterior cerebral infarction, sparing memory-related structures in the diencephalon, may represent a means of investigating the role of hippocampal afferents and the parahippocampal areas in memory processing. Among 30 patients with unilateral posterior cerebral infarction a group of 12 subjects with left-sided lesions suffered from marked verbal memory and learning dysfunction, whereas the remaining subjects with left or right-sided lesions showed no obvious memory deficit. The impairment was most prominent for verbal learning tasks, while recall of isolated and complex verbal information appeared to be less affected. In some cases the memory disturbance could be detected up to one year after the causative event. Analysis by CT scanning revealed a coincidence of mnestic disturbances with lesions around the left collateral sulcus affecting the posterior parahippocampal gyrus and the collateral isthmus. The latter term refers to the fibre stem of the medial temporal lobe limited by the floor of the lateral ventricle and the depth of the collateral sulcus. Through this bottle-neck, bidirectional fibres run between the posterior parahippocampal gyrus and different sensory-specific and multimodal association areas. A lesion within the collateral isthmus or, more particularly, a combined lesion of the collateral isthmus and the posterior parahippocampal gyrus, deprives the hippocampus itself of its main afferent projection source and, on the other hand, prevents dissemination of the hippocampal output to widespread neocortical areas. It cannot be excluded, at least in some cases, that small lesions in the fimbria-fornix route, in the retrosplenial cortex or in the hippocampal formation itself also contribute to the memory and learning dysfunction. However, these small lesions, so far as they could be detected by CT scanning, were present in patients with and without memory disturbances.