Omission mismatch negativity of speech sounds reveals a functionally impaired temporal window of integration in schizophrenia

Highlights

• Mismatch negativity patterns in patients with schizophrenia differed from those in controls.

• Impaired temporal windows of integration may cause mismatch negativity patterns.

• Our data supported biomarker-based diagnosis and reclassification of schizophrenia.

Abstract

Objective

We hypothesized that sensory memory associated with the temporal window of integration (TWI) would be impaired in patients with schizophrenia, an issue that had not been evaluated using omission mismatch negativity (MMN) of complex speech sounds. We aimed to assess the functional changes in auditory sensory memory associated with the TWI in patients with schizophrenia by investigating the effect of omission of complex speech stimuli on the MMN.

Methods

In total, 17 patients with schizophrenia and 15 control individuals participated in the study. The MMN in response to omission deviants of complex speech sounds was recorded, while the participants were instructed to ignore the series of speech sounds.

Results

The MMN latency in patients with schizophrenia was significantly prolonged by deviant stimuli to omissions corresponding to the early and late parts of the temporal TWI. There were no significant group differences in the amplitude of the MMN to omissions at different time points across the TWI.

Conclusions

Our results suggested that sensory tracing function in patients with schizophrenia is impaired in the early and the later half of the TWI.

Significance

We showed that certain MMN abnormalities in patients with schizophrenia may be caused by an impaired TWI.

Introduction

Humans can detect changes in sounds automatically. This phenomenon first reported by Näätänen et al. (1992) who measured an event-related potential named mismatch negativity (MMN). The MMN is elicited by an infrequent change in homogenous repetitive stimuli, even in the absence of attention. The major MMN sources are the auditory cortex, including Heschl’s and superior temporal gyrus (Hari et al., 1984, Scherg et al., 1989). The MMN has been measured to evaluate the mechanisms underpinning mental disease in terms of cognitive function. It is generated by sensing changes relative to auditory sensory memory (Näätänen et al., 1978). Thus, auditory sensory memory could play a role by transiently keeping the previous sound-image for comparison with each incoming sound. The existence of a transient memory naturally suggests that successive acoustic information should be divided into a unitary image with a given time contrast due to its memory size. Such a unitary event storage was considered to be related to the brain mechanism, the so-called temporal window of integration (TWI), which integrates the closely presented sounds into single information units.

Several studies on the TWI have been reported to date. Tervaniemi et al. (1994) reported that the omitted second tone elicited the MMN only when the offset-onset interval between the paired tones was 40 or 140 ms, but not when this interval was 240 or 340 ms. This finding supported the temporal integration hypothesis that sensory memory traces integrate auditory information concerning closely spaced stimuli into a unitary event. Yabe et al., 1997, Yabe et al., 1998 conducted experiments to determine the exact TWI duration using a paradigm in which a click sound was used as a standard, and its omission was deviant while various stimulus onset asynchrony (SOA) in different blocks. Their results suggested that the duration of the temporal window was 160–170 ms, as a definite MMN was elicited with a constant SOA ranging from 70 to 160 ms, whereas the MMN was not elicited with SOA > 170 ms. This finding indicated that the memory trace of sound-image should be represented as time-functional information from 0 to 170 ms. Additionally, Yabe et al., 2001, Yabe et al., 2004 assessed the effects of omission deviants in complex tone sound stimuli corresponding to the TWI on the MMN, and reported that the amplitude decreased and the latency was prolonged, as omissions were located in late parts of the stimuli. They also reported that the difference in the relative latency to omission indicated the accuracy of the tracing function in sensory memory. Furthermore, discriminative sensitivity was not uniform over the TWI duration because their MMN responses declined nonlinearly toward the temporal window terminal. Moreover, this scientific team (2001) also elicited the MMN with the same peak latency but with a different peak amplitude using two types of deviant stimuli differed from each other only in the period after the silent segment occurrence. The mismatch process obtained by their paradigm was triggered at the moment of change but still lasted after detection of deviation. This finding suggested that standard and deviant stimuli were treated as a unitary event within a TWI. Especially, the TWI integrates neighboring sounds and/or silent into a single percept.

A recent research in healthy control participants using omission in complex speech sounds suggested that the MMN latency was prolonged and the amplitude decreased, when omissions were located in late parts of stimuli (Hoshino et al. 2020, unpublished data, in submission). This implies that the automatic auditory discrimination function in the TWI is most robustly strengthened at approximately 50 ms, and discriminative capacity decreases toward the latter half of the TWI. In psychiatry, the attenuation of MMN amplitude in patients with schizophrenia has been reported in many studies, and is suggestive of dysfunction in the preattentive processing of auditory information (Shelley et al., 1991). In particular, the duration MMN (dMMN) has a more significant decrease in amplitude (Todd et al., 2008). Moreover, the dMMN abnormality has a greater detection sensitivity for schizophrenia compared to the frequency and intensity MMN because of the potential functional impairment of the TWI in patients with schizophrenia. Especially, impairments in the dMMN may be enhanced owing to greater deficits in the latter half of the TWI. Further, after studying the sound effects of the MMN in patients with schizophrenia, Kasai et al. demonstrated a language-related dysfunction at the early stage of auditory processing of speech sounds by comparing the MMN of tone and speech sounds (Kasai et al., 2002, Kasai et al., 2003). Therefore, their results suggested that speech sounds were more appropriate in investigating the characteristics of schizophrenia pathophysiology compared to the tone sounds.

Based on these findings, we considered that a paradigm corresponding to the TWI, composed of speech sound, might identify the functional changes in auditory sensory memory associated with the TWI in patients with schizophrenia. We hypothesized that the MMN corresponding to the latter part of the TWI may be different in such patients compared with healthy controls, when assessed by measuring the MMN to omitted segments in complex speech sounds. Therefore, the present study aimed to identify the functional changes in auditory sensory memory associated with the TWI in patients with schizophrenia by investigating the effects of omission in complex speech stimuli on the MMN.