Auditory verbal hallucinations (AVHs) are characterized as the experiences of hearing a voice or any other sounds for languages under the conditions of conscious state with no external stimulus (McCarthy-Jones and Resnick, 2014; Upthegrove et al., 2016). As the most prominent and burdensome symptoms of schizophrenia (SCZ), the prevalence of AVH reported rates of 60–80% (Alderson-Day et al., 2015). However, roughly 30% of patients with AVH [treatment-resistant auditory verbal hallucinations, (TRAVHs)] are not responded well for current clinical treatments (Kennedy et al., 2014) and might cause an increased risk of violence, suicide, and hospitalization (Upthegrove et al., 2016; Dugré et al., 2018). From 2013 to 2018, the International Consortium on Hallucination Research (ICHR) suggested that clinical and pathological features of AVH should be explored from all perspectives (Waters, 2012; Jardri et al., 2019).

Currently, the underlying pathological mechanisms remain poorly understood, but an array of studies have explored the neural correlations of AVH. For instance, the functional neuroimaging evidence explored the various brain areas related to AVH, some of which were involved in inner speech generation, inner speech perception, and hallucinations processing (Stephane et al., 2001; Jardri et al., 2011; Kühn and Gallinat, 2012). According to cognitive neurophysiology, inner speech has played an important role in a series of theories of AVH in patients with SCZ (Shergill, 2003; Langland-Hassan, 2010; Wayne, 2012). One prominent hypothesis suggests that inner speech fails to be labeled as internally generated, and instead being perceived as an autonomous, non-self-voice (Jones, 2010). The neurological evidence of inner speech model involved in left inferior frontal gyrus, bilateral temporal cortex, the supplementary motor area, premotor cortex, precentral, postcentral gyri, inferior parietal lobe, right insula, and posterior cerebellar cortex bilaterally (Moseley et al., 2013; Curčić-Blake et al., 2017). Other results displayed that AVH may arise from the abnormal activation in the left superior temporal region and the left prefrontal cortex during the processing of inner speech (Allen et al., 2007; Simons et al., 2010; Jardri et al., 2011). These above research displayed the neural activation associated with AVH and inner speech were found to offer more convincing evidence for the disorder of inner speech in patients with AVH. Based on these findings, we hypothesized that functional abnormalities in the prefrontal and temporal cortex, as measured with functional near-infrared spectroscopy (fNIRS), underly TRAVH among patients with SCZ, supplying more evidence to have a better understanding of the inner speech model of AVH.

Functional near-infrared spectroscopy (fNIRS) (Herrmann et al., 2003) is a non-invasive imaging device to evaluate brain function, which has several advantages. fNIRS is easy to set up, requires minimal, constrains, and does not occupy a large space (Wei et al., 2020). Additionally, it is safe, nonrestrictive, quiet, and economical for measurement (Ferrari and Quaresima, 2012). fNIRS signals are believed to assess the neuronal activity according to a phenomenon known as neurovascular coupling (Gsell et al., 2000) and have strong correlations of blood oxygenation level-dependent (BOLD) signals measured by fMRI (Sasai et al., 2012). Moreover, fNIRS is less sensitive to motion artifacts and can be operated without the acoustic scanner noise, which is more suitable for evaluating patients with AVH during related tasks (Ehlis et al., 2014). fNIRS can monitor both oxy- and deoxy-hemoglobin levels in the cerebral cortex. Since the oxy changes are greater than deoxy-hemoglobin, oxy-hemoglobin is used as the brain activity marker (Ferrari and Quaresima, 2012), while the verbal fluency task (VFT) is considered as the sensitive indicators of language function and frontal and temporal lobe function, such as multiple high-level cognitive process, including executive function, attention, inhibition, and working memory, which has been proposed as an assessment measure of functional impairment and illness severity (Herrmann et al., 2003; Sasai et al., 2012; Koike et al., 2013). In addition, a large number of fNIRS studies have explored and compared the prefrontal cortex activation between patients with SCZ and healthy control (HC) by VFT. Previous findings showed that patients with SCZ had reduced oxy-hemoglobin in the bilateral prefrontal and temporal cortices during the VFT task (Suto et al., 2004; Takizawa et al., 2008, 2014; Koike et al., 2013, 2017; Kumar et al., 2017), which indicated that VFT is a widely accepted cognitive task for elucidating brain dysfunction related to psychotic state. Among patients with SCZ, the symptom of AVH was associated with the decreases in the cerebral blood flow in the brain regions, including postcentral gyri (inner speech production) (Tremblay et al., 2003) and right supplementary motor area (inner speech imagery) (McGuire et al., 1995; Raij and Riekki, 2012), as well as the bilateral dorsolateral prefrontal cortices (inner speech monitoring) (Cui et al., 2017). The gray matter density reduction in right postcentral gyri was found to be specifically negatively related to the severity of AVH(García-Martí et al., 2008). Whereas, such studies were argued to favor that patients with SCZ with AVH and without AVH showed significantly decreased ReHo in the bilateral postcentral gyrus (Zhuo et al., 2016), based on the different results, it is even more important to know whether the AVH-related brain region, such as postcentral gyri, is the common neural substance underlying both the inner speech and AVH. Specifically, for SCZ-TRAVH, clarification of neural mechanisms is of great clinical significance for treatment. Therefore, in the present study, we utilized multi-channel fNIRS to investigate whether the symptom of TRAVH is associated with functional alterations in the brains of patients with SCZ during VFT and the underlying neurological mechanisms for the inner speech related to AVH.

This study used fNIRS to explore the association between hemodynamic response in the prefrontal and temporal cortices and TRAVH symptom in patients with SCZ. Our results indicated that patients with SCZ-TRAVH exhibited lower activation in the right postcentral gyrus than did those without the symptom of AVH.

First, in terms of oxy-hemoglobin, our study observed that the SCZ-TRAVH and SCZ-nAVH group had lower brain activation in extensive prefrontal and temporal cortices in contrast to the HC group, which were similar to the previous fNIRS studies showing the lower oxy-hemoglobin response in similar brain cortices during the related task (Takizawa et al., 2008, 2014; Wei et al., 2020). In addition, the brain activity of 24 channels in the SCZ-TRAVH group was lower than that in the HC group, whereas the oxy-hemodynamic response of only 18 channels in the SCZ-nAVH group was lower. According to the previous fMRI studies, hypoactivation of the bilateral prefrontal cortex and temporal cortices in patients with SCZ has been reported, which is associated with confusion of imagined items as perceptual events (Simons et al., 2006). Taken together, the results suggested that hemodynamic dysfunction in the prefrontal cortex and temporal cortices may be more extensive in patients with SCZ-TRAVH than patients without AVH.

In our result, SCZ-TRAVH group have lower hemodynamic response at channel 22 of the right postcentral gyrus compared to SCZ-nAVH group. This result was replicated in the fMRI study that patients with AVH showed decreased cerebral blood flow in the bilateral superior and postcentral gyri compared with patients with SCZ-nAVH (Cui et al., 2017). In addition, a significant negative association emerged between the Δβ value of oxy-hemoglobin in the region of ch22 and severity of AVH symptom [PSYRATS-AH score, PANSS-positive score, and P3 (hallucinations) score]. The association between Δβ values of channel 22 and PSYRATS-AH remained significant after controlling for other scales scores and VFT performance in a multiple regression analysis. This meaningful finding partially agreed with previous researches reported significant association between the reduction of bilateral postcentral gyrus volume and the severity AVH in SCZ (García-Martí et al., 2008; Nenadic et al., 2010). It was not difficult to find that among the above associated scales, PSYRATS-AH had the highest correlation, followed by PANSS-positive score. According to the consensus, PSYRATS-AH is a more detailed and multidimensional examination than PANSS and P3 for hearing voices, which had generally strong reliability and been widely used in diverse clinical studies of neural correlations of AVH (Ratcliff et al., 2011). Considering the highest correlation of PSYRATS-AH with CH22 signal changes, this may largely suggested associations of decreased activation in the postcentral gyrus with the putative mechanisms by which psychotic symptoms are generated, especially AVH. Additionally, the previous fMRI studies reported the right postcentral gyrus related to AVH and may be a new therapeutic target in the future (van Lutterveld et al., 2013). Additionally, Zui Narita et al.'s research further explored the potential utility of fNIRS for the prediction of effects of tDCS on psychotic symptoms (Narita et al., 2018). Although the findings show different potential therapeutic brain regions, these results support the concept that fNIRS may provide a valid method to evaluate brain functions based on blood flow, similar to fMRI (Lee et al., 2018).

The mechanism by which brain activity measured by fNIRS should be considered in relation to suffering from AVH symptoms. The postcentral gyrus regarded as one location of classically and primary defined somatosensory cortex (Skipper et al., 2005; Pérez-Bellido et al., 2018). A meta-analysis reflected the association of AVH with activation in the postcentral regions related to subvocal speech (Kühn and Gallinat, 2012) and therefore being considered involving inner speech generation (Feinberg, 1978; Shergill et al., 2002), while the symptoms of AVH were consistently associated with the disorders of generation of inner speech (inner verbal thoughts) (McGuire et al., 1996; DeLisi, 2001). According to other neuroimaging researches, inner speech-related brain region involved in the left inferior frontal, the right pre- and postcentral, and both superior temporal gyri (Shergill et al., 2002). Besides, another recent study further demonstrated that the postcentral gyrus was involved in the auditory information process and even auditory hallucination symptoms (Joo et al., 2020). It seemed that the inner speech generation played a role in the pathogenesis of AVH (McGuire et al., 1995). Taken together with the above neurological evidence, the decreased activation of postcentral gyrus may underlie the basis for disorders of inner speech generation and imply alterations in the monitoring of inner speech, leading to a higher risk for suffering from the symptom of AVH, partially consistent with inner speech model of AVH.

In this study, we focus on a limited group of SCZ-TRAVH whose AVH symptom was at a relatively stable phase to develop in-depth knowledge of pathological features of AVH and the heterogeneous nature of this disease. Additionally, we used fNIRS to reveal abnormal brain function during VFT task, which was corresponded to the resting-state and task-state fMRI studies' results (van Lutterveld et al., 2013; Cui et al., 2017), which make fNIRS as a suitable research tool to explore the underlying neural alterations of patients with TRAVH during related tasks. Our significant finding may promote the development in the design of the novel therapeutic strategies for TRAVH and encourage the application of fNIRS analysis to follow-up studies with a larger sample size to explore the underlying neuroimaging mechanism and understand the pathophysiological correlations of TRAVHs in SCZ.

Nevertheless, there are some clear limitations associated with our research. First, patient's sample sizes in this study were relatively small for aiming to offer a specific activation pattern for SCZ-TRAVH. Therefore, increasing the sample size might identify further brain regions associated with AVH. Second, although our researchers focused on studying a limited group of patients with SCZ with stable psycho-pathological features (AVH) resistant to drug treatment to obtain the activation difference in potential brain regions, we cannot fully rule out the medicinal effects. However, according to Koike et al.'s study (Koike et al., 2013), they emphasized that there is no association between fNIRS activity of patients with SCZ and doses of drugs. Furthermore, there was no significant difference in medication treatment between the SCZ-TRAVH and SCZ-nAVH groups in our study. In addition, we did not obtain data related to cognitive symptoms in our participants. Since cognitive symptoms are associated with the prefrontal dysfunction for patients with SCZ, further study with an assessment of cognitive symptoms is warranted.

The present multi-channel fNIRS study showed an association between TRAVH-induced hemodynamic impairments in the right postcentral gyrus and a history of TRAVH among patients with SCZ. These findings may help to provide evidence for the common theories regarding the “inner speech” and AVH, facilitating to do the design of novel therapeutic methods. A further study is remained to explore the long-term outcomes of SCZ-TRAVH through the fNIRS measurement.

Our study used NIRS technology to explore and compare the abnormal activation difference in brain regions for patients with SCZ with TRAVH and nAVH. We found a negative association between the decreased activation of right postcentral gyrus and severity of TRAVH. Further, this study supplied a possible explanation for the occurrence of TRAVH in SCZ from the inner speech insights, which may have implications for an early biological evidence-based intervention for TRAVH.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Requests to access these datasets should be directed to 761620286@qq.com.

Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

NL, YT, and SL contributed to the study design and were involved in data acquisition, analysis, and interpretation. XL, DW, QL, and YX contributed to the study design and data interpretation. All authors participated in the drafting or critical review of the article, gave final approval of the version to be published, and agreed to be accountable for all aspects of the work.

This work was supported by the National Natural Science Foundation of China (81971601 and 81571319), the National Key Research and Development Program of China (2016YFC1307004), the Wu Jieping Foundation (320.6750.18283), the special project for guiding the transformation of scientific and technological achievements in Shanxi Province (201904D131020), and the Multidisciplinary Team for Cognitive Impairment of Shanxi Science and Technology Innovation Training Team (201705D131027).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.