Functional retrograde amnesia refers to a loss of remote memories in the absence of structural brain damage. In this study, functional retrograde amnesia is comprehensively investigated across five patients (A.B., C.D., E.F., G.H., and I.J.). Investigation of the patients' histories, neuropsychological and psychological-psychiatric evaluation was carried out. Furthermore, in two functional neuroimaging experiments brain correlates were compared within and across patients while they attempted to retrieve autobiographical events prior and following the onset of the amnesia.

Standard neuropsychological investigation included tests of intelligence, anterograde memory, attention, executive functions, and Theory of Mind (ToM) abilities. Current psychological-psychiatric status and personality were assessed by clinical screenings and questionnaires. Assessment of retrograde memory performance comprised tests of semantic and episodic remote memory. A functional magnetic resonance imaging (fMRI) study was conducted with three of the patients and a positron emission tomography (PET) experiment enrolled two patients. In both experiments, patients were given visual presentations of descriptions of personal episodes originating from the time preceding or following the amnesia onset. In the fMRI experiment, fictitious old and fictitious new episodes were additionally included. Data were analysed contrasting single conditions.

The patients showed minor impairment across neuropsychological domains attention/executive functions, anterograde memory and ToM. Psychological screening revealed that most patients showed psychiatric symptoms, and personality deviations in the sense of social desirability. In remote memory testing, all patients except A.B. exhibited difficulties in parts of semantic remote memory. Tests of autobiographical remote memory revealed that A.B. was selectively amnesic for episodic information. Patients C.D. and I.J. were entirely amnesic for their lives prior to the onset of amnesia. Patients E.F. and G.H. were amnesic for the prior 12-14 years only. In the neuroimaging experiments, comparisons of true memory conditions showed that retrieval of episodes from the amnesic time period were associated with enhanced activation in right prefrontal brain regions in patients E.F. and I.J. and left lateral superior temporal/inferior parietal regions in patient A.B. Processing of new in contrast to old episodes revealed enhanced activation in posterior association cortex in three patients (E.F., G.H., I.J.). In processing of true and fictitious episodes within both time periods, activations did not differ between attempts to retrieve true compared to fictitious episodes in the amnesic time period. Concerning the non-amnesic time, however, significant differential activations associated with true relative to fictitious episodes were seen in A.B. and C.D., as well as a non-significant trend in the same direction in patient E.F.

The discussion concentrates on probable organic and psychological mechanisms contributing to the onset and maintenance of the amnesia. Psychological factors were present in the majority of the patients. The most obvious indicators were fugue-like state (I.J.), preceding depression (A.B.), stressful current life events (A.B., G.H.) or childhood trauma (G.H.), previous transient memory loss (C.D., G.H.), previous psychiatric background (C.D., G.H.) and current somatic complaints of psychogenic origin (A.B.). Some of the latter factors and behavioural anomalies are alternatively interpreted as signs of malingering. The most definite evidence for organic factors could be gathered in patient E.F. who exhibited hypometabolism of brain glucose in resting state PET, temporally graded retrograde amnesia without additional psychiatric background, stress, or obvious secondary gain.

The right prefrontal activity during access of old memories observed in patients E.F. and I.J. may reflect enhanced retrieval effort or episodic retrieval mode within this time period. The left lateralised activation of lateral superior temporal/inferior parietal areas in patient A.B. points to enhanced processing of language. Processing of new memories in contrast to old memories was associated with activation in posterior association cortex (E.F., G.H., I.J.); this may indicate enhanced processing of sensory-perceptual features of to-be-retrieved memories in the non-amnesic time. Significant differential activations between true versus fictitious events were restricted to the non-amnesic time. These were related to memory- or emotion-related brain areas in A.B. and C.D. and may also have been present in E.F. Therefore, only within the accessible time period neural activity differentiated true from fictitious episodes which was clearly paralleled by behaviour.