Cognitive dissonance resolution depends on episodic memory.

The notion that past choices affect preferences is one of the most influential concepts of social psychology since its first report in the 50 s, and its theorization within the cognitive dissonance framework. In the free-choice paradigm (FCP) after choosing between two similarly rated items, subjects reevaluate chosen items as more attractive and rejected items as less attractive. However the relations prevailing between episodic memory and choice-induced preference change (CIPC) remain highly debated: is this phenomenon dependent or independent from memory of past choices? We solve this theoretical debate by demonstrating that CIPC occurs exclusively for items which were correctly remembered as chosen or rejected during the choice stage. We used a combination of fMRI and intra-cranial electrophysiological recordings to reveal a modulation of left hippocampus activity, a hub of episodic memory retrieval, immediately before the occurrence of CIPC during item reevaluation. Finally, we show that contrarily to a previous influential report flawed by a statistical artifact, this phenomenon is absent in amnesic patients for forgotten items. These results demonstrate the dependence of cognitive dissonance on conscious episodic memory. This link between current preferences and previous choices suggests a homeostatic function of this regulative process, aiming at preserving subjective coherence.

Probing ERP correlates of verbal semantic processing in patients with impaired consciousness.

Our ability to identify covert cognitive abilities in non-communicating patients is of prime importance to improve diagnosis, to guide therapeutic decisions and to better predict their cognitive outcome. In the present study, we used a basic and rigorous paradigm contrasting pairs of words orthogonally. This paradigm enables the probing of semantic processing by comparing neural activity elicited by similar words delivered in various combinations. We describe the respective timing, topography and estimated cortical sources of two successive event-related potentials (ERP) components (N400 and late positive component (LPC)) using high-density EEG in conscious controls (N=20) and in minimally conscious (MCS; N=15) and vegetative states (VS; N=15) patients recorded at bedside. Whereas N400-like ERP components could be observed in the VS, MCS and conscious groups, only MCS and conscious groups showed a LPC response, suggesting that this late effect could be a potential specific marker of conscious semantic processing. This result is coherent with recent findings disentangling early and local non-conscious responses (e.g.: MMN in odd-ball paradigms, N400 in semantic violation paradigms) from late, distributed and conscious responses (e.g.: P3b to auditory rule violation) in controls and in patients with disorders of consciousness. However, N400 and LPC responses were not easily observed at the individual level, - even in conscious controls - , with standard ERP analyses, which is a limiting factor for its clinical use. Of potential interest, the only 3 patients presenting both significant N400 and LPC effects were MCS, and 2 of them regained consciousness and functional language abilities.