Skip to main content

Hippocampus and memory in children with developmental amnesia

Final Activity Report Summary - HM (Hippocampus and memory in children with developmental amnesia.)

Developmental amnesia (DA), resulting from early damage to the hippocampus in the medial temporal lobe, is characterised by chronic difficulties with memory for everyday events, i.e. with episodic memory. Patients with DA acquire nevertheless normal intelligence, academic skills and factual knowledge, indicating relatively spared semantic memory. To account for this preserved ability, we proposed that cognitive memory was organised hierarchically, both functionally and anatomically. According to this model, context-free semantic memory depended mainly on neocortical and parahippocampal interactions, whilst context-rich episodic memory required the subsequent step of parahippocampal and hippocampal interaction. Although this proposed hierarchy was consistent with functional neuroanatomical evidence in the monkey, further research with patients was needed to assess its validity. We addressed two questions derived from this model.

Firstly, the spontaneous ability to distinguish novel objects from familiar ones, i.e. incidental recognition memory, was impaired in monkeys with early or late hippocampal damage. This same ability was also impaired in humans with adult onset amnesia. Whether the impairment in patients was due to selective damage to the hippocampus remained to be determined. To explore this, patients with DA, who suffered from selective hippocampal pathology and controls were examined on the visual paired comparison task. In this task, pairs of identical pictures of landscapes or patterns were presented on a screen and, after a variable delay, one of the previously presented pictures was shown again, but this time paired with a novel one. Participants were instructed to simply look at the pictures while their eye fixations were recorded. Preferential looking at the new picture signalled novelty detection and the recognition of the previously presented picture as ‘familiar’. Results showed that DA patients’ preferential looking to novel pictures was as good as controls at zero to five seconds delay between familiarisation and recognition, but significantly impaired at delays of 30 or 120 sec. Incidental recognition memory was therefore impaired only in cases when information had to be stored beyond 30 sec, suggesting that the integrity of the hippocampus was required for delayed incidental recognition in humans.

The second question was related to the hypothesis that the relatively spared semantic memory of patients with DA was due to the integrity of the cortical regions adjacent to the hippocampus, i.e. the temporal pole, entorhinal, perirhinal and posterior parahippocampal cortices, areas which were thought to be important for acquiring semantic knowledge. The segmentation and volumetric measurement of these parahippocampal areas in brain scans depended upon the use of hippocampal landmarks. However, the reduction of hippocampal volume caused by episodes of hypoxia or ischaemia in DA hindered the use of the hippocampal landmarks in segmentation and volumetric measurement, because these references could have changed due to shrinkage of the hippocampus. Therefore, the aim of the second study was to use extra-hippocampal landmarks, identifiable in magnetic resonance images (MRIs), for volumetric measurement of the cortical areas comprising the parahippocampal region. Results showed that the criteria for the identification of adult parahippocampal regions in MRI could also be applied to the healthy developing brain. Furthermore, the lateral geniculate nucleus, situated outside the medial temporal lobe, could be used as an extra-hippocampal landmark for the boundary between the entorhinal and posterior parahippocampal cortices. It was, however, necessary to validate the use of the lateral geniculate nucleus as a landmark in MRI scans obtained from patients with DA. This called for further analysis of the structural MRI scans of patients with DA to determine the relationship between the shrunken hippocampus, the lateral geniculate nucleus and the entorhinal or posterior parahippocampal boundary.