An fMRI study of multimodal selective attention in schizophrenia

Andrew R. Mayer, J. Christopher Edgar, Terri Teshiba, Stefan D. Klimaj, Josef M. Ling, Andrew B. Dodd, Vince D. Calhoun, Juan Bustillo, Trent ToulouseThe British Journal of Psychiatry 2015 · DOI

Abstract

Results suggest a prominent role for dysfunction within auditory, sensorimotor and parietal areas relative to prefrontal CCN nodes during multisensory cognitive control.

Introduction

Cognitive control—the capacity to prioritize goal-relevant information while suppressing distractors—is fundamentally impaired in schizophrenia. Historically, research has attributed these deficits to "top-down" dysfunction within the Cognitive Control Network (CCN), primarily localized in the prefrontal cortex. However, real-world cognitive demands often involve complex multisensory environments. This study investigates whether the primary bottleneck in schizophrenia lies within the "CEO" of the brain (the CCN) or at the level of sensory processing ("bottom-up" regulation).

Methods

Researchers utilized a multisensory Stroop task during fMRI scanning to evaluate 33 clinically stable outpatients with schizophrenia and 33 matched healthy controls. Participants were cued to either "HEAR" or "LOOK" at streams of numbers (1, 2, or 3) presented simultaneously via audio and video.

Experimental task design (A, B) and behavioral reaction times (C, D) for healthy controls (HC) and schizophrenia patients (SP).

As illustrated in the figure above, conflict was introduced through incongruent trials (e.g., audio "two" vs. video "three"). Cognitive load was manipulated by varying stimulus frequency (0.33 Hz vs. 0.66 Hz). The study measured Attention Related Modulations (ARMs) to assess how effectively participants upregulated sensory cortex activity in response to increased task demands.

Results

Behavioral Performance

Patients with schizophrenia exhibited significantly slower reaction times and reduced accuracy compared to healthy controls. This performance gap was most pronounced during high-frequency, incongruent trials. In the auditory task, patients averaged approximately 650ms compared to 595ms for controls, a trend visible in the reaction time distributions (Panels C and D).

fMRI Findings: The Prefrontal Surprise

Contrary to the prevailing "top-down" hypothesis, the study found no significant group differences in the activation of prefrontal CCN nodes. Effect sizes for these differences were remarkably small (Cohen’s $d$ between $0.03$ and $-0.22$), suggesting that the core executive machinery was functioning similarly across both groups.

Sensory Hyperactivation and Dysregulation

Instead of prefrontal deficits, patients showed significant hyperactivation in the right secondary auditory cortex (BAs 13/43) and bilateral sensorimotor/parietal regions. This suggests that the brains of patients were working harder but less efficiently at the initial stages of information processing.

Furthermore, a critical breakdown was observed in sensory "tuning." While healthy controls showed robust positive modulation (upregulation) of the primary auditory cortex as task difficulty increased ($t_{32} = 3.15, P < 0.05$), this mechanism was entirely absent in the patient group.

Discussion

These findings challenge the traditional focus on prefrontal dysfunction in schizophrenia. The results suggest that in multisensory environments, the primary deficit may be "bottom-up," characterized by sensory cortex hyperactivation and a failure to dynamically allocate resources to sensory processing under load.

This shift in perspective has significant implications for therapeutic interventions. If cognitive deficits are driven by sensory processing inefficiencies rather than executive strategy failures, then cognitive remediation should perhaps prioritize "bottom-up" sensory training. By improving the fidelity of sensory input, we may alleviate the burden on higher-order networks, ultimately improving functional outcomes for individuals living with schizophrenia.

Limitations and Future Directions

The study focused on clinically stable, medicated outpatients; results may vary in acute or unmedicated populations. Additionally, the inherent noise of the MRI scanner may have influenced auditory task performance. Future research should explore whether these sensory tuning deficits can serve as biomarkers for functional outcomes or as targets for sensory-based cognitive remediation therapies.

Summary updated

Abstract

Results suggest a prominent role for dysfunction within auditory, sensorimotor and parietal areas relative to prefrontal CCN nodes during multisensory cognitive control.
BibTeX
@article{mayer2015fmri,
  author = {Andrew R. Mayer and J. Christopher Edgar and Terri Teshiba and Stefan D. Klimaj and Josef M. Ling and Andrew B. Dodd and Vince D. Calhoun and Juan Bustillo and Trent Toulouse},
  title = {{An fMRI study of multimodal selective attention in schizophrenia}},
  journal = {The British Journal of Psychiatry},
  year = {2015},
  doi = {10.1192/bjp.bp.114.155499},
}
Figure