Reduced communication between frontal and temporal lobes during talking in schizophrenia

Abstract

Background: Communication between the frontal lobes, where speech and verbal thoughts are generated, and the temporal lobes, where they are perceived, may occur through the action of a corollary discharge. Its dysfunction may underlie failure to recognize inner speech as self-generated and account for auditory hallucinations in schizophrenia.

Methods: Electroencephalogram was recorded from 10 healthy adults and 12 patients with schizophrenia (DSM-IV) in two conditions: talking aloud and listening to their own played-back speech. Event-related electroencephalogram coherence to acoustic stimuli presented during both conditions was calculated between frontal and temporal pairs, for delta, theta, alpha, beta, and gamma frequency bands.

Results: Talking produced greater coherence than listening between frontal-temporal regions in all frequency bands; however, in the lower frequencies (delta and theta), there were significant interactions of group and condition. This finding revealed that patients failed to show an increase in coherence during talking, especially over the speech production and speech reception areas of the left hemisphere, and especially in patients prone to hallucinate.

Conclusions: Reduced fronto-temporal functional connectivity may contribute to the misattribution of inner thoughts to external voices in schizophrenia.

Introduction

Connection between frontal and temporal cortical areas via the arcuate fasciculus provides a pathway by which frontal speech production areas can prime the auditory cortex for impending speech Creutzfeldt et al 1989, Paus et al 1996b, Petrides and Pandya 1988. Dysfunction of this connection might be responsible for “failure to integrate perception and action,” Friston and Frith 1995, Weinberger and Lipska 1995 “splitting of mental faculties,” (Bleuler 1911/1950) and may underlie some of the positive symptoms of schizophrenia.

Priming of the auditory cortex by the frontal lobes may occur through the action of a “corollary discharge.” Although originally described in the visual system as a mechanism to control flight in the horsefly, corollary discharge has been used to describe a mechanism in the auditory system that allows self-monitoring of the spoken word. Such a mechanism might serve speech acquisition during infancy and our ability to distinguish between our own and others’ utterances throughout life. Thus, corollary discharge might be a key mechanism for monitoring our own speech, thoughts, and behaviors. An example of corollary discharge relevant to speech perception is the inhibition of auditory cortical responsiveness during phonation. In subhuman primates, about 50% of call-responsive neurons, identified by response to prerecorded cells, are inhibited during phonation (Muller-Preuss and Ploog 1981). In humans, preoperative intracranial recordings show a reduction in responsiveness of neurons in the middle temporal gyrus, starting 100 msec before and continuing during the patient’s speech, but not occurring when another person speaks to the patient (Creutzfeldt et al 1989).

Disrupted connectivity between frontal and temporal lobes (Friston and Frith 1995) is consistent with the hypothesis that a defective corollary discharge mechanism in schizophrenia causes the misperception of thoughts as voices Feinberg 1978, Feinberg and Guazzelli 1999, Frith 1995. While corollary discharges are hypothesized to accompany inner speech and thoughts, our most complex motor acts (Jackson 1958), their effects on auditory cortex should be particularly evident during speech production which, normally involves differential perception and processing of self-generated auditory input such as vocalizations. Accordingly, we assessed fronto-temporal functional connectivity in controls and patients with schizophrenia during speaking aloud relative to simply hearing playback of speech. The speech statements employed were typical in content and form to those reported by patients experiencing auditory hallucinations (Nayani and David 1996).

Connectivity between brain regions can be assessed using various brain imaging techniques. Those based on hemodynamic responses (e.g., functional magnetic resonance imaging [fMRI] and positron emission tomography [PET]) provide relatively good spatial but poor temporal resolution of cortical activity. Those based on electrophysiological responses (e.g., electroencephalogram [EEG] and Magnetoencephalogram [MEG]) provide limited spatial resolution, but excellent (e.g., millisecond) temporal resolution. Using fMRI or PET, investigators may infer functional relationships between disparate cortical areas coactivated by the same task conditions (Fletcher et al 1998), but little can be concluded about the relative time course of these activations or the degree to which their neural activity is temporally coupled. Using EEG, investigators may calculate the relationship between electrophysiological signals recorded from two different cortical areas and make inferences about their interconnection. This can be done using a variety of techniques including cross-covariance calculations (Gevins et al 1987), wavelet transformations (Nikolaev et al 2001), and coherence calculations (Thatcher et al 1986), the most common of which is coherence.

Coherence is a frequency-dependent measure of the degree of relatedness between EEG recorded over two different brain areas. High coherence between two brain areas indicates that their amplitudes at a given frequency and their associated phase angles are correlated across time epochs (Lachaux et al 1999). When coherence is low, it indicates that across time epochs, the relationship between power in the two signals and/or the relative phase difference between them is inconsistent. Moreover, the coherence measure does not allow specification of whether the relationship between the two signals is stronger in terms of relative phase or power (Lachaux et al 1999). Coherence can range from zero to one, and it does not distinguish positive from negative correlations between frequency amplitudes across time. Accordingly, it can reflect either inhibition or excitation of connected areas (Manganotti et al 1998). In any case, high levels of coherence between EEG recorded from noncontiguous electrodes reflect interdependence between brain regions (Lachaux et al 1999) due to anatomical connections (Fein et al 1988), functional coupling (Thatcher et al 1986), “perceptual binding,” (Gray and Singer 1989) and/or associative learning (Miltner et al 1999).

Electroencephalogram coherence can be calculated in passive, unstimulated, task-free conditions and also during cognitively engaging tasks. Patients with schizophrenia show decreases in delta and theta coherence over the left frontal lobes in a task-free condition (Tauscher et al 1998), reduction of bilateral anterior activation during frontal lobe tasks (Morrison-Stewart et al 1996), a reduced level of focal activation of left frontal areas during left hemisphere activation tasks (Morrison-Stewart et al 1991), and reductions in frontal-temporal alpha band coherence in patients with reality distortion during performance of a mathematical task (Norman et al 1997).

Electroencephalogram coherence can also be calculated when subjects are processing external stimuli, from which event-related brain potentials (ERPs) can be derived. A recent study (Muller et al 1999) reported increased delta and theta coherence when normal subjects were processing names compared with nouns. Using event-related coherence analysis, investigators (Rappelsberger et al 1994) have reported increased event-related alpha band coherence between premotor and motor areas. Furthermore, event-related coherence is useful in attempting to understand the temporal relationships between averaged ERPs recorded over different cortical areas (Leocani et al 1997). To date, studies of event-related EEG coherence have not been reported in patients with schizophrenia.

To understand the temporal relationships between averaged ERPs to acoustic probes during speaking, we calculated event-related coherence from single trials recorded over frontal and temporal areas. Using these same data, we have previously shown that the averaged ERP N1 component to acoustic probes is differentially sensitive to schizophrenia during talking and listening (Ford et al 2001c). To understand whether this reflects a reduction in temporal communication between frontal and temporal areas, we assess the temporal synchrony of single-trial EEG recorded over frontal and temporal cortical areas by calculating the coherence of the EEG signals. Our predictions were that theta would show increased fronto-temporal coherence during speaking relative to listening to speech in controls, but not in patients. Furthermore, this lack of coherence during speech in patients would be most evident in those most prone to auditory hallucinations. Such a finding would reflect normal functional connectivity between speech production and perception areas during speech and its disruption in schizophrenia. We also expected these effects to be stronger over the left than right hemisphere, and strongest over lateral frontal and posterior temporal areas involved in speech production and reception, respectively.