Objectives: Anecdotal reports suggest that following traumatic brain injury (TBI) retrograde memories are initially impaired and recover in order of remoteness. However, there has been limited empirical research investigating whether a negative gradient in retrograde amnesia—relative preservation of remote over recent memory—exists during post-traumatic amnesia (PTA) compared with the acute phase post-emergence. This study used a repeated-measures design to examine the pattern of personal semantic (PS) memory performance during PTA and within two weeks of emergence to improve understanding of the nature of the memory deficit during PTA and its relationship with recovery. Methods: Twenty patients with moderate-severe TBI and 20 healthy controls (HCs) were administered the Personal Semantic Schedule of the Autobiographical Memory Interview. The TBI group was assessed once during PTA and post-emergence. Analysis of variance was used to compare the gradient across lifetime periods during PTA relative to post-emergence, and between groups. Results: PS memory was significantly lower during PTA than post-emergence from PTA, with no relative preservation of remote memories. The TBI group was still impaired relative to HCs following emergence from PTA. Lower overall PS memory scores during PTA were associated with increased days to emerge from PTA post-interview. Conclusions: These results suggest a global impairment in PS memory across lifetime periods particularly during PTA, but still present within 2 weeks of emergence from PTA. PS memory performance may be sensitive to the diffuse nature of TBI and may, therefore, function as a clinically valuable indicator of the likely time to emerge from PTA. (JINS, 2018, 24, 1064–1072)

With time and experience, memories undergo a process of reorganization that involves different neuronal networks, known as systems consolidation. The traditional view, as articulated in standard consolidation theory (SCT), is that (episodic and semantic) memories initially depend on the hippocampus, but eventually become consolidated in their original forms in other brain regions. In this study, we review the main principles of SCT and report evidence from the neuropsychological literature that would not be predicted by this theory. By comparison, the evidence supports an alternative account, the transformation hypothesis, whose central premise is that changes in neural representation in systems consolidation are accompanied by corresponding changes in the nature of the memory. According to this view, hippocampally dependent, episodic, or context-specific memories transform into semantic or gist-like versions that are represented in extra-hippocampal structures. To the extent that episodic memories are retained, they will continue to require the hippocampus, but the hippocampus is not needed for the retrieval of semantic memories. The transformation hypothesis emphasizes the dynamic nature of memory, as well as the underlying functional and neural interactions that must be taken into account in a comprehensive theory of memory. (JINS, 2011, 17, 766–780)

To assess if amnesics have intact remote memory for general semantic information, we examined memory for vocabulary words with known dates of entry into the language between 1955 and 1989. Amnesics of mixed etiology with acute onset performed normally on both a recall and a recognition task. Korsakoff patients, in contrast, were impaired on both tasks and demonstrated a gradient such that their knowledge of words acquired during more recent time periods was worse than that of words acquired during more remote time periods. The improvement in performance associated with recognition testing was larger for Korsakoff patients than for control subjects and correlated significantly with a composite measure of frontal dysfunction. These findings suggest a deficit in the controlled search and retrieval of semantic information in Korsakoff patients. The implications of the differential performance of Korsakoff and mixed etiology amnesics for explanations of temporally graded retrograde amnesia are discussed. (JINS, 1995, I, 443–453.)

Dans le domaine des troubles de la mémoire induits par les benzodiazépines, un des résultats le plus communément admis est que l’information acquise avant la prise du produit est conservée intacte. A l’opposé, les troubles de la mémoire antérograde ne sont pas rares. L’intensité et la durée des troubles de la mémoire antérograde semblent dépendre des facteurs suivants: nature de la benzodiazépine utilisée, dose administrée, voie d’administration, épreuve de mémoire utilisée, nombre d’heures après la prise auxquelles les informations sont présentées et les épreuves de mémoire sont employées, caractéristiques des populations testées. Les épreuves impliquant en quelques secondes une remémoration de peu d’items (jusqu’à 6) ne sont pas perturbées par les benzodiazépines. De la même façon, lorsqu’il s’agit de listes de mots le rappel des derniers mots de la liste n’est généralement pas affecté par les benzodiazépines, tandis que les mots de milieu de liste sont plus difficilement remémorés. En général quand les études ont exploré la mémoire différée, les effets amnésiogènes se sont révélés être plus importants que lorsque la mémoire est testée immédiatement après apprentissage. Deux principales explications des effets des benzodiazépines sur la mémoire ont pu être proposées. Il semble que l’hypothèse d’un défaut de consolidation recueille plus d’arguments en sa faveur que celle d'un apprentissage «état-dépendant». La sédation contribue probablement aux effets amnésiogènes mais ne les explique pas.