EEG CORRELATES OF DREAM-RECALL IN DEPRESSED OUTPATIENTS AND HEALTHY CONTROLS

The present study explored EEG correlates of dream recall in 17 sympto matic, unmedicated depressed patients and in 19 healthy adults. EEG se gments from the last 30 minutes of sleep, from the five minutes follow ing morning awakening, and the absolute difference between sleep and w aking EEG were contrasted between the two groups of participants durin g successful dream recall and during no recall. Period amplitude analy sis was used to quantify EEG frequencies. Increased high-frequency bet a incidence in the right hemisphere and amplitude in both hemispheres during sleep were associated with dream recall in both patient and con trol groups. Depressed patients also showed higher delta amplitude in both hemispheres during sleep associated with recall, hut this effect did not reach significance. With regard to the changes between sleep a nd wakefulness, a smaller change in right hemisphere beta and delta in cidence characterized successful recall in healthy controls. By contra st, those with depression showed recall success when the sleep/wake sh ifts in right hemisphere beta and delta incidence were larger. Recall failure was characterized by small EEG shifts from sleep to wakefulnes s in the depressed group. The same effects were observed for beta and delta amplitude measures, except that healthy controls showed a large shift in delta amplitude in the sleep-wake transition during successfu l recall but not during recall failure. Recall in those with depressio n was associated with a dramatic shift in left hemisphere delta amplit ude. These findings provide support for Koukkou and Lehmann's (1983, 1 993) state-shift hypothesis of dream recall in healthy controls (excep t for left hemisphere delta amplitude) but not in the depressed. It ap pears that in order to recall a dream, depressed patients must undergo larger shifts lit brain activity and perhaps a different pattern of r eorganization of EEG frequencies than controls. This finding may accou nt for the low rates of recall reported previously in this clinical gr oup.