Source monitoring and proneness to auditory-verbal hallucinations: a signal detection analysis

It has been suggested that a bias to misattribute self-generated thoughts to a nonself source underlies the experience of auditory-verbal hallucinations (AVH). We tested this hypothesis with healthy participants prone or not prone to AVH.

Participants (N=133) were presented with 96 words for subsequent recognition (half positively, half negatively valenced). For self-generated trials, participants generated a sentence containing the word. For other-generated trials, participants heard a prerecorded sentence containing the word. At test, studied words were re-presented visually, intermixed with 96 matched lures. Participants indicated the study status (old or new) and source (self or other) for each item. Sensitivity and bias measures were derived for item and source memory using signal detection theory. The 20 participants scoring highest on questions relating to AVH from the revised Launay-Slade Hallucination Scale formed the high-AVH group and the 20 scoring lowest formed the low-AVH group.

ANOVAs revealed no significant differences between the two participant groups in sensitivity or bias of source memory, regardless of item valence. There was a trend for the sensitivity of item memory to be lower in the high-AVH group, compared with the low-AVH group. The bias of item memory was not significantly different between groups.

Although we found no evidence that source-monitoring problems underlie the experience of AVHs in the general population, we recommend that signal detection measures be applied in future investigations of source monitoring in at-risk and clinical populations.