Neurocognitive bases of self-monitoring of inner speech in hallucination prone individuals

Abstract

Verbal hallucinations in schizophrenia patients might be seen as internal verbal productions mistaken for perceptions as a result of over-salient inner speech and/or defective self-monitoring processes. Similar cognitive mechanisms might underpin verbal hallucination proneness in the general population. We investigated, in a non-clinical sample, the cerebral activity associated with verbal hallucinatory predisposition during false recognition of familiar words -assumed to stem from poor monitoring of inner speech-vs. uncommon words. Thirty-seven healthy participants underwent a verbal recognition task. High- and low-frequency words were presented outside the scanner. In the scanner, the participants were then required to recognize the target words among equivalent distractors. Results showed that verbal hallucination proneness was associated with higher rates of false recognition of high-frequency words. It was further associated with activation of language and decisional brain areas during false recognitions of low-, but not high-, frequency words, and with activation of a recollective brain area during correct recognitions of low-, but not high-, frequency words. The increased tendency to report familiar words as targets, along with a lack of activation of the language, recollective, and decisional brain areas necessary for their judgement, suggests failure in the self-monitoring of inner speech in verbal hallucination-prone individuals.

Figure 1

Activation clusters positively associated with verbal hallucination proneness score during correct recognitions (yellow) and false recognitions (red) of low-frequency words, while controlling for visual imagery score, sex, verbal IQ, and Pr-global. The slices depicting MNI coordinates derive from peak activations for the low-frequency word contrasts reported in Tables 1 and 2.

Figure 2

Activation clusters negatively associated with visual imagery score during correct recognitions (purple) and false recognitions (green) of high-frequency words, while controlling for verbal hallucination proneness score, sex, verbal IQ, and Pr-global. The slices depicting MNI coordinates derive from peak activations for the high-frequency word contrasts reported in Tables 1 and 2.