Paracingulate Sulcus Length Is Shorter in Voice-Hearers Regardless of Need for Care

Abstract

Hallucinations—while often considered an indication of mental illness—are commonly experienced by those without a need for clinical care. These nonclinical voice-hearers offer an opportunity to investigate hallucinations in the absence of confounds inherent to the clinical state. Recent work demonstrates an association between auditory verbal hallucinations (AVH) and structural variability in paracingulate sulcus (PCS) of medial prefrontal cortex in a clinical population. However, before PCS length may be considered a biomarker for clinical hallucination risk, it is necessary to investigate PCS structure in a nonclinical population of voice-hearers with AVH phenomenology similar to those of their clinical counterparts. In the current study, PCS length was measured from T1-weighted structural MRI scans of four groups of participants: (1) voice-hearers with a psychotic disorder (n = 15); (2) voice-hearers without a psychotic disorder (n = 15); (3) nonvoice-hearers with a psychotic disorder (n = 14); and (4) nonvoice-hearers without a psychotic disorder (n = 15). There was a main effect of AVH status—but not psychosis—on right PCS length, with no interaction of AVH and psychosis. Participants with AVH exhibited reduced right PCS length compared to participants without AVH (mean reduction = 8.8 mm, P < 0.05). While past studies have demonstrated decreased PCS length in clinical voice-hearers, ours is the first demonstration that shorter right PCS extends to nonclinical voice-hearers. Our findings support the hypothesis that differences in PCS length are related to the propensity to hear voices and not to illness, consistent with a continuum model of voice-hearing.

Introduction

While hallucinations are often considered to be the sine qua non of serious mental illness, they are commonly experienced by those without a need for care.¹ Such nonclinical hallucinating individuals offer a window into the etiology of hallucinations. Studying a nonclinical population of voice-hearers provides an opportunity to investigate auditory verbal hallucinations (AVHs) in the absence of confounds such as stigma, distress, or medication, among other factors integral to the clinical state.

Recent work highlights a potential contribution of cortical morphology to the etiology of AVH. Garrison et al assessed structural variability in paracingulate sulcus (PCS) of medial prefrontal cortex. They found that a reduction of 10 mm in left hemisphere PCS length increased the likelihood of AVH by 19.9%.² However, these findings were in clinical hallucinators, and reduced leftward asymmetry in PCS has been repeatedly demonstrated in schizophrenia and associated with reduced reality monitoring.³ Similar analyses have been conducted in nonclinical voice-hearers and in those participants, no reduction in PCS length was reported.⁴

There is considerable phenomenological variability in the voice-hearing experiences of nonclinical voice-hearers. Put simply, some may be more clinical than others. We recently identified a relatively homogenous group of nonclinical voice-hearers who identify as clairaudient psychic mediums. Their voice-hearing experiences are very similar to those reported by a matched patient voice-hearing group⁵ and their brain and behavioral responses to a perceptual inference task were, at least at the level of prior beliefs, similarly perturbed to participants who hear voices within the context of psychosis.⁶ Despite similar AVH symptom severity, this group of nonclinical voice-hearers differed from their clinical counterparts in their reports of less distress and more volitional control over the voices.

These observations are consistent with a continuum model of voice-hearing,^1,5,7–13 along which this sample of voice-hearers may be nearer to patients with schizophrenia in some respects than others. By contrast, the PCS data reported by Garrison and colleagues may cause some to question whether such a continuum could exist: if a single morphological feature such as PCS length can distinguish between clinical and nonclinical voice-hearers, perhaps they are distinct populations separated by a discrete neural feature.⁴ However, in comparison to the clairaudient nonclinical voice-hearers we identified, their nonclinical voice-hearing group experienced relatively infrequent voices.¹⁴ Therefore, their nonclinical group not only differed from their clinical counterparts in psychosis status, but also in AVH phenomenology. We sought to determine if the differences in PCS length also applied to the clairaudient nonclinical voice-hearers.

If PCS length is to be developed as a biomarker for clinical hallucination risk, then it is important to demonstrate that it is not similarly reduced in nonclinical voice-hearers.

Methods

Recruitment, clinical assessment, and image acquisition are detailed in Powers et al. T1-weighted structural MRI scans were obtained using 3D MPRAGE sequence from a 3-Tesla Siemens magnetic resonance imaging (MRI) scanner for 4 groups of participants: (1) voice-hearers with a psychotic disorder (P+H+; n = 15); (2) voice-hearers without a psychotic disorder (P−H+; n = 15); (3) nonvoice-hearers with a psychotic disorder (P+H−; n = 14); and (4) nonvoice-hearers without a psychotic disorder (P−H−; n = 15). All analyses were carried out blind to participant group.

Full details regarding recruitment and clinical characterization of this sample have been published elsewhere.⁶ In brief, voice-hearing severity was quantified using 2 clinician-rated scales (the PANSS [P3 Item] and the Brief Psychosis Rating Scale), as well as 2 self-report scales (the LSHS-R and the Auditory Hallucination Rating Scale [AHRS]). Nonclinical voice-hearers (ie, the self-identified clairaudient psychic mediums) were required to be antipsychotic-naive and not in treatment of any psychiatric issue, and upon further screening could not have a diagnosable Axis I disorder; no participants were excluded after interview for this reason. Finally, the majority (69%) of recruited nonvoice-hearing participants with a psychotic disorder had no prior voice-hearing experiences. Of those who did, a mean of 8 years had elapsed since their last voice-hearing experience.

Scans were imported into and calibrated using Mango brain visualization software (version 3.6). Using sagittal slices, 4 mm to the left or right of the medial line, the PCS was identified and measured following prior work.² All analyses were conducted in MATLAB 2014b.

Results

With 2-way ANOVA, there was a main effect of AVH status—but not psychosis—on right PCS length, with no interaction of AVH and psychosis (figure 1a). This effect was not seen in the left PCS (figure 1b). Participants with AVH exhibited reduced right PCS length compared with participants without AVH (mean reduction = 8.8 mm, P < .05). The effect of AVH on right PCS length was sufficiently robust to persist with the addition to the model of the potentially confounding covariates: age and Intelligence Quotient (IQ) (F = 4.55, P < .05); right cortical surface area (F = 4.40, P < .05); right hemisphere gyrification index (F = 4.39, P < .05); and intracranial volume (F = 4.24, P < .05). While right PCS length is numerically smaller in nonclinical voice-hearers (P−H+), this difference was not statistically significant.

Fig. 1.

Paracingulate sulcus (PCS) length. Right PCS is significantly shorter in participants with hallucinations than in participants without (a), but this effect is not seen in left PCS (b).

Discussion

While past studies have demonstrated decreased PCS length in individuals with psychosis and AVH compared with their nonhallucinating counterparts,² ours is the first demonstration that such an effect extends to nontreatment-seeking voice-hearers. Right PCS in our sample was significantly shorter in voice-hearers, regardless of whether or not they had a diagnosable psychotic illness. In fact, mean right PCS length was numerically shorter for nonclinical voice-hearers than for any other group, although this difference did not reach statistical significance. Our findings support the hypothesis that differences in PCS length are specifically related to the propensity to hear voices and not to illness, disability, or dysfunction.^2,15 However, our nonclinical voice-hearers’ experiences were very similar to those reported by patients with schizophrenia⁵: while voice-hearing severity was remarkably similar between the 2 groups with hallucinations, there were a few key differences: the nonclinical voice-hearing participants were less distressed by their experiences and reported more volitional control over the onset and offset of the voices than the clinical group. Because PCS length is not different between these groups, PCS length is unlikely to underpin these phenomenological differences and therefore should not be considered a predictor of need for care in voice-hearing populations. However, it is possible that not all nonclinical voice-hearers have experiences—and, we would argue—brains—that match patients so closely.

While our findings demonstrate shorter right PCS in voice-hearers, regardless of psychotic illness, previous studies report an association of shorter right and left PCS with hallucinations only in the context of a psychotic disorder.⁴ The extension of shorter right PCS to our nonclinical voice-hearers may reflect phenomenological differences between the studies: our nonclinical voice-hearers were required to have daily AVH, whereas the Garrison cohort needed only have monthly AVH (as detailed in Sommer et al¹⁴). Frequency of voice-hearing may be driving the differences in findings. In light of these discrepancies, future work should aim to disentangle the more nuanced phenomenological characteristics that could covary with PCS length.

In addition, the lack of group differences in the left PCS observed in our study may reflect a parallel lack of differences in reality monitoring between our groups—if PCS is needed for reality monitoring as previously suggested.^3,16 While reality monitoring was not directly measured in our study, clinical and nonclinical groups did not differ in their sense of reality of AVH. It is possible that not all nonclinical voice-hearers are equal in their reality monitoring, which may additionally explain differences in PCS length findings between our cohort and that of Garrison and colleagues.

The present findings are unable to address the fundamental question of how PCS length or associated cortical morphological changes may translate to fundamental perceptual processes thought to underpin hallucinations. This determination may depend on methods from computational neuroscience, in part because biologically informed computational models hold potential for describing normal and pathological information processing within a common framework.^17–21 Using related methods, we recently demonstrated a hallucination-specific effect in an adjacent prefrontal region during conditioned hallucinations.⁶ We showed that this region’s activity encodes low-level perceptual beliefs about the sensory environment that may drive hallucinations. It may be that the morphological differences highlighted here play a role in mediating perceptual belief formation or its interactions with incoming sensory evidence during hallucinations. Recently identified methods for functional MRI at the resolution of cortical layers (ie, laminar functional MRI²²) may aid in fully characterizing these links between PCS morphology and the computational underpinnings of hallucinatory phenomena.

Funding

ARP was supported by a NARSAD Young Investigator Award from the Brain and Behavior Research Foundation, a K23 Career Development Award from the National Institute of Mental Health (NIMH; K23 MH115252-01A1), by a Career Award for Medical Scientists from the Burroughs Wellcome Fund, by NIMH R21MH122940, and by the Yale Department of Psychiatry and the Yale School of Medicine. PRC was supported by the Yale University Department of Psychiatry, the Connecticut Mental Health Center (CMHC) and State of Connecticut Department of Mental Health and Addiction Services (DMHAS), an IMHRO/Janssen Rising Star Translational Research Award, NIMH R01MH12887, NIMH R01MH120089, and NIMH R21MH116258. JRG was supported by a University of Cambridge Parke Davis Exchange Fellowship in Biomedical Sciences.