Heschl’s gyrus (HG), a convolution on the superior temporal plane, hosts the primary auditory cortex (Rademacher et al., 1993; Da Costa et al., 2011) and is also involved in memory (Weinberger, 2015) and emotional (Concina et al., 2019) processing. The morphology of HG markedly varies across individuals, with approximately 30–50% of healthy individuals potentially having complete or partial duplication (Leonard et al., 1998; Rademacher et al., 2001; Abdul-Kareem and Sluming, 2008; Marie et al., 2015). This anatomical variant appears to reflect variations in cytoarchitectonic development during gestation (Chi et al., 1977; Armstrong et al., 1995), and duplicated HG may lead to learning disabilities (Leonard et al., 1993, 2001) and reduced HG activity during auditory processing (Tzourio-Mazoyer et al., 2015) in a non-clinical population. In a recent magnetic resonance imaging (MRI) study, we reported an increased prevalence of HG duplications in first-episode schizophrenia (Sz) (Takahashi et al., in submission), which may reflect the early neurodevelopmental pathology (Weinberger, 1987; Insel, 2010). However, since another MRI study that specifically examined HG duplication patterns in chronic Sz did not find significant results (Hubl et al., 2010), it currently remains unclear whether illness stages affect the HG pattern of Sz. Furthermore, although structural/functional abnormalities in the superior temporal plane may underlie the positive psychotic symptoms (Alderson-Day et al., 2015; Takahashi and Suzuki, 2018) as well as core trait abnormalities [e.g., deficits in social cognition (Mier et al., 2017) and verbal fluency (Antonova et al., 2004)] of Sz, it has not yet been clarified whether the HG gyrification pattern is associated with these clinical features.

MRI studies on individuals at high risk of developing psychosis [i.e., at-risk mental state (ARMS) (Yung et al., 2004, 2005)], who have an increased risk of developing psychosis within a short period of time [approximately 30% at 2 years (Fusar-Poli et al., 2012a)], generally showed similar gross morphological characteristics associated with early neurodevelopment [e.g., an altered sulcogyral pattern in the orbitofrontal region (Nakamura et al., 2019) and widespread cortical hypergyria (Sasabayashi et al., 2017)] to those of overt Sz. Since these brain anomalies are at least partly observed in participants without a later onset of psychosis (Sasabayashi et al., 2017; Nakamura et al., 2019), they may represent biological traits associated with general vulnerability to psychopathology. These gross brain characteristics may contribute to cognitive impairments in the Sz and ARMS groups (Takahashi et al., 2019a), supporting the presentation of cognitive impairments, particularly in social function (Lee et al., 2015) and verbal fluency (Fusar-Poli et al., 2012b), even before the onset of psychosis as a trait vulnerability marker. However, despite evidence of partly shared superior temporal gray matter reductions in the ARMS and Sz groups (Takahashi et al., 2010b), no MRI studies to date have specifically examined the HG duplication pattern and its potential contribution to clinical features (e.g., cognitive deficits) in the ARMS cohort.

Therefore, the present MRI study aimed to examine the HG gyrification pattern (single HG, partial duplication, and complete duplication) in ARMS individuals and Sz patients, compare it with those in healthy controls, and examine its potential contribution to clinical variables (symptoms, social and cognitive functions). Based on our previous MRI findings from an independent sample of Sz (Takahashi et al., in submission) as well as the potential role of brain gyrification as a stable neurodevelopmental marker (Chi et al., 1977; Armstrong et al., 1995), we predicted increased HG duplication in both the ARMS and Sz groups. We also speculated that the HG pattern in these groups may be associated with clinical variables that reflect trait abnormalities, such as cognitive impairments.

To the best of our knowledge, this is the first MRI study to examine the HG duplication pattern in clinical high-risk individuals for developing psychosis. We demonstrated that ARMS individuals and patients with established Sz both exhibited a significantly higher prevalence of duplicated HG patterns than healthy controls. Furthermore, the HG pattern was associated with global symptom ratings and verbal fluency ability in these participants. The present results suggest that the gross morphological characteristics of the superior temporal plane represent vulnerability factors associated with psychosis, which may be associated with clinical trait abnormalities.

The present study replicated our previous findings from an independent cohort of first-episode Sz (Takahashi et al., in submission) showing increased HG duplication in Sz patients and also demonstrated that illness stages (i.e., first-episode vs. chronic stages) did not significantly influence HG patterns. On the other hand, a previous study by Hubl et al. (2010) only found a slightly higher prevalence of duplicated HG in chronic Sz patients. However, their negative finding may be partly due to the small sample size examined (13 Sz and 13 control participants) as well as their definition of HG duplication, which classified the CSD pattern as a variant of single HG. Since we demonstrated increased HG duplication in Sz regardless of the subtype (i.e., CSD or CPD), the Sz group examined by Hubl et al. (2010) must have had a higher prevalence of the duplicated HG pattern according to the traditional HG pattern definition [single vs. duplicated (CSD or CPD) (Leonard et al., 1998; Rademacher et al., 2001; Abdul-Kareem and Sluming, 2008; Marie et al., 2015)]. While the mechanisms regulating the development of cortical gyrification remain unclear, the secondary gyri of HG, which form variations in the HG gyrification pattern, predominantly develop during the late gestation period (i.e., after 36 weeks of gestation) (Chi et al., 1977) along with local neuronal connectivity and synaptic development (Van Essen, 1997), but remain stable after birth (Armstrong et al., 1995). Therefore, HG gyrification studies in Sz generally support the notion that the gyrification pattern in Sz represents a stable trait marker associated with early neurodevelopmental pathology (Matsuda and Ohi, 2018).

One of the primary results of the present study was that ARMS individuals, who may be vulnerable to psychopathology but will not necessarily develop overt psychosis (Yung et al., 2004; Fusar-Poli et al., 2012a), exhibited an increased HG duplication pattern similar to that in Sz. Based on the potential relationship between brain gyrification and local neuronal connectivity (Van Essen, 1997), the present results appear to be consistent with previous functional neuroimaging findings showing that the ARMS and Sz groups share local connectivity disruption involved in HG (Yoon et al., 2015; Du et al., 2018). A few MRI studies on cortical surface features in clinical high-risk individuals also showed similar gross morphological characteristics, such as altered sulcogyral patterns (Sasabayashi et al., 2017; Nakamura et al., 2019) and sulcal-depth abnormalities (Takahashi et al., 2019b), with patients with established Sz. In contrast to the evidence of active gray matter reductions in the superior temporal plane (e.g., HG and planum temporale) during the early illness stages of psychosis (Takahashi and Suzuki, 2018), a recent longitudinal study demonstrated the stability of gyrification features during the clinical high-risk period as a marker of early neurodevelopmental insults (Damme et al., 2019). Nevertheless, high-risk individuals with the later onset of psychosis may exhibit greater gyrification abnormalities before illness onset (Sasabayashi et al., 2017; Das et al., 2018) because greater and/or more prolonged neurodevelopmental deviations during gestation and consequent anomalous post-pubertal brain changes may lead to overt and sustained psychosis (Pantelis et al., 2005). Since the present ARMS group with a short follow-up period (median = 2.4 years) only examined a small number of participants with a later onset of psychosis (N = 5), the potential of the HG gyrification pattern as a predictive marker of the later onset of psychosis remains unclear.

The present results suggested that the partial duplication of HG (i.e., CSD) was associated with a more severe general psychopathology in ARMS individuals, supporting aberrant connectivity in the superior temporal region potentially contributing to prodromal-like symptoms (Yoon et al., 2015). However, the present Sz cohort (predominantly chronic cases) did not replicate the relationship between the CPD pattern and mild positive symptom severity observed in first-episode Sz (N = 62) (Takahashi et al., in submission), implicating that neurodevelopmental pathology may be associated with susceptibility to positive psychotic symptoms of Sz but this relationship may be influenced by various factors including illness stages and treatment. On the other hand, as also suggested in our sample (Table 1), cognitive deficits, particularly in verbal fluency and memory functioning, may exist even before the onset of psychosis as markers of increased vulnerability (Fusar-Poli et al., 2012b; Lee et al., 2015). In the present study, we found that participants with the left CSD pattern had a greater deficit in verbal fluency, but not in other domains or social functioning, than those with the left CPD pattern in the Sz (N = 63) or combined Sz and ARMS (N = 120) groups. This result appears to be consistent with the notion that candidate neural circuits for verbal fluency deficits include the superior temporal region for both the Sz (Frith et al., 1995; Antonova et al., 2004) and ARMS (Meijer et al., 2011) groups. While the functional role of the HG duplication type (i.e., CPD vs. CSD) remains largely unknown, participants with the CSD pattern may have a significantly smaller planum temporale gray matter than those with the CPD pattern bilaterally for both the Sz and control groups (Takahashi et al., in submission), which may lead to deficits in verbal ability (Shapleske et al., 1999). However, the potential contribution of different HG patterns to the pathophysiology of psychotic disorders warrants further study at various illness stages, particularly using functional neuroimaging.

Several potential limitations in the present study need to be addressed. First, as described above, it was not possible to examine whether the HG gyrification pattern was associated with the future onset of psychosis because only 5 participants (8.8%) in the ARMS group developed psychosis in the clinical follow-up period. Furthermore, the ARMS group was younger than the other groups in the present study. Second, the majority of Sz and 14 ARMS participants were being treated with antipsychotics during the present study. These factors were not expected to significantly affect gross sulcogyral patterns; however, antipsychotic medication may be a confounding factor for the morphology of the superior temporal plane (Takahashi and Suzuki, 2018) and cognitive functioning (Keefe, 2014). Therefore, future studies using a larger antipsychotic naïve ARMS cohort (particularly participants with a later onset of psychosis) and well-matched comparison groups are needed to examine the HG gyrification pattern and its potential contribution to clinical features (including the later onset of psychosis). Third, we did not correct our results of ANOVA/ANCOVA for multiple comparisons due to exploratory nature of our study. We predicted that the HG pattern would be associated with cognitive impairments, but we had no clear hypothesis and comprehensively assessed the potential contribution of HG pattern to all available cognitive subdomains, which might lead to potential Type I error. Finally, since superior temporal gray matter reductions (Takahashi et al., 2010a, c) and altered brain gyrification patterns (Yang et al., 2016; Maggioni et al., 2019) have been reported in other neuropsychiatric disorders (e.g., mood and anxiety disorders and autism), the disease specificity of the present results warrant further study.

The results of this MRI study demonstrated that clinical high-risk individuals for psychosis exhibited an increased HG duplication similar to that in patients with Sz, which may reflect common vulnerability factors. These groups partly shared cognitive impairments, which were associated with HG gyrification patterns. We also found a relationship between the HG pattern and severity of general symptoms observed in high-risk individuals. Therefore, the gross morphology of the superior temporal plane may represent the biological trait abnormalities of Sz that exist prior to illness onset; however, our findings should be replicated in an independent and larger cohort especially for high-risk individuals with and without the later onset of psychosis in order to investigate potential role of HG pattern as a predictive marker of Sz.

The data analyzed in this study is subject to the following licenses/restrictions: the datasets generated during the current study will not be available for public use, since we do not have permission to share the data. Requests to access these datasets should be directed to TT, tsutomu@med.u-toyama.ac.

The studies involving human participants were reviewed and approved by the Committee on Medical Ethics of Toyama University. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

MS, YH, and TT conceived the idea and methodology of the study. TT conducted the statistical analyses and wrote the manuscript. DS, YH, MK, and HK recruited participants and were involved in clinical and diagnostic assessments. TT, DS, and TP analyzed MRI data. YM and SN assessed the sociocognitive functions of the study participants. KN provided technical support for MRI scanning and data processing. AF managed the MRI and clinical data. MS and YT contributed to the writing and editing of the manuscript. All authors contributed to and approved the final manuscript.

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.