To contribute to the identification of brain regions involved in déjà-vu, we studied the metabolic pattern of cortical involvement in patients with seizures of temporal lobe origin presenting with or without déjà-vu. Using voxel-based analysis of 18FDG-PET brain scans, we compared glucose metabolic rate of 8 patients with déjà-vu, 8 patients without déjà-vu, and 20 age-matched healthy subjects. Patients were selected after comprehensive non-invasive presurgical evaluation, including normal brain MRI and surface electroclinical features compatible with unilateral temporal lobe epilepsy (TLE).
Patients with and without déjà-vu did not differ in terms of age, gender, epilepsy lateralization, epilepsy onset, epilepsy duration, and other subjective ictal manifestations. TLE patients with déjà-vu exhibited ipsilateral hypometabolism of superior temporal gyrus and of parahippocampal region, in the vicinity of perirhinal/entorhinal cortex, in comparison either to healthy subjects or to TLE patients without déjà-vu (p < 0.05 FDR-corrected). By contrast, no difference was found between patient subgroups for hypometabolism of hippocampus and amygdala. At an individual-level, in comparison to healthy subjects, hypometabolism of both parahippocampal region and superior temporal gyrus was present in 7/8 patients with déjà-vu. Hippocampal metabolism was spared in 3 of these 7 patients.
Déjà-vu is experienced by most people, occurring about once a year in healthy subjects (Brown, 2003). Its frequency decreases with age, and it appears to be promoted by stress and fatigue. Authors have associated this phenomenon to drowsiness or to possible ictal events in normal subjects (Spatt, 2002). On the other hand, déjà-vu is classically reported with similar characteristics by some patients during seizures of temporal lobe origin, suggesting that déjà-vu functionally involves the same temporal structures in epileptic patients and healthy subjects (Bancaud et al., 1994, Gloor, 1990, Spatt, 2002 and Wild, 2005).
Ictal déjà-vu was described for the first time by Hughlings Jackson within the spectrum of the “dreamy state”, which also included vivid reminiscences of memories. Dreamy state was supposed to be linked to lesions affecting the medial temporal lobe (MTL) (Jackson, 1888 and Jackson and Stewart, 1899). Later, Penfield and Jasper associated déjà-vu with sensory illusions and emotional disturbances as experiential hallucinations induced by stimulation of lateral temporal neocortex, particularly the superior temporal gyrus (Penfield & Jasper, 1954). More recently, Bancaud et al. showed that dreamy state depends upon electrical activity of a neural network involving both medial and lateral temporal lobe areas (Bancaud et al., 1994). Among MTL structures, many authors have emphasized the role of the amygdala, the hippocampus (Gloor et al., 1982 and Vignal et al., 2007) and more recently rhinal cortices (Barbeau et al., 2005a and Bartolomei et al., 2004). However, the exact contribution of each of these structures in déjà-vu and dreamy state remains debated (Spatt, 2002).