Altered Small-World Brain Networks in Schizophrenia Patients during Working Memory Performance

Hao He, Jing Sui, Qingbao Yu, Jessica A. Turner, Beng‐Choon Ho, Scott R. Sponheim, Dara S. Manoach, Vincent P. Clark, Vince D. CalhounPLoS 2012 · DOI

Introduction

Schizophrenia (SZ) is characterized by significant impairments in working memory (WM), the cognitive system responsible for the temporary maintenance and manipulation of information. While previous research has identified localized brain abnormalities in SZ, particularly within the prefrontal cortex, the broader organizational principles of the whole-brain network remain less understood. This study utilizes graph theory to examine the "small-world" architecture of the brain—a topology that optimizes for both local specialized processing (clustering) and global information integration (short path lengths). We hypothesized that this efficient organization would be disrupted in SZ patients as cognitive demands increase.

Methods

The study analyzed fMRI data from 35 patients with chronic schizophrenia and 35 demographically matched healthy controls (HC) from the Mind Clinical Imaging Consortium (MCIC). Participants performed the Sternberg Item Recognition Paradigm (SIRP) at three difficulty levels: Low (1 digit), Medium (3 digits), and High (5 digits).

Functional nodes were defined using Group Independent Component Analysis (ICA) to identify 105 task-related spatial blocks. Edges between these nodes were constructed using partial correlations of BOLD signal time courses, a method that minimizes the influence of indirect dependencies. The resulting networks were thresholded to maintain a small-world regime and analyzed using several key metrics:

  • Clustering Coefficient ($C_{net}$): Measures local connectivity and specialization.
  • Local Efficiency ($E_{local}$): Reflects the network's fault tolerance and local information exchange.
  • Global Efficiency ($E_{global}$): Indicates the capacity for parallel information processing.
  • Characteristic Path Length ($L_{net}$): Represents the average distance between any two nodes.

Results

Behavioral Performance

As expected, SZ patients demonstrated significantly lower accuracy and longer reaction times compared to HC across all memory loads ($p < 0.01$).

Network Topology and Load Effects

The most significant group differences emerged at the medium WM load. At this level, SZ patients exhibited significantly lower clustering coefficients and reduced local efficiency compared to HC.

A critical finding was the difference in network stability. Healthy controls maintained stable network metrics across all three load levels, suggesting a robust neural architecture. In contrast, SZ patients showed high volatility; their networks became significantly less clustered and less locally efficient as the load increased from low to medium. Interestingly, a slight "rebound" in these metrics was observed at the high load in the SZ group, potentially indicating a non-linear compensatory response to extreme cognitive demand.

Brain-Behavior Correlations

The altered network structure in SZ patients was directly linked to their task performance. Specifically, lower clustering and lower local efficiency at the medium load were significantly correlated with longer reaction times, suggesting that the breakdown of local specialized processing directly contributes to cognitive slowing.

Discussion

These findings suggest that the schizophrenic brain becomes "diffuse" under moderate cognitive pressure. While healthy brains maintain a tightly-knit, clustered organization, SZ networks lose this local specialization, leading to inefficient information processing. This "diffuse" state may occasionally result in shorter global path lengths, but it lacks the specialized local "cliques" necessary for effective working memory performance.

Limitations

The study acknowledges several limitations, including the potential confounding effects of antipsychotic medication, although no direct correlation with symptoms was found. Additionally, the use of binarized, undirected networks may overlook nuances in connection strength and information flow directionality. Future research using weighted and directed graphs (e.g., via Granger Causality) is recommended to further elucidate these complex dynamics.

Conclusion

By demonstrating that small-world brain architecture in schizophrenia is load-dependent and inefficient, this study provides a structural framework for understanding the cognitive bottlenecks observed in the disorder. The breakdown of local network efficiency at moderate task demands appears to be a hallmark of the schizophrenic brain's response to cognitive load.

Summary updated

Abstract

Impairment of working memory (WM) performance in schizophrenia patients (SZ) is well-established. Compared to healthy controls (HC), SZ patients show aberrant blood oxygen level dependent (BOLD) activations and disrupted functional connectivity during WM performance. In this study, we examined the small-world network metrics computed from functional magnetic resonance imaging (fMRI) data collected as 35 HC and 35 SZ performed a Sternberg Item Recognition Paradigm (SIRP) at three WM load levels. Functional connectivity networks were built by calculating the partial correlation on preprocessed time courses of BOLD signal between task-related brain regions of interest (ROIs) defined by group independent component analysis (ICA). The networks were then thresholded within the small-world regime, resulting in undirected binarized small-world networks at different working memory loads. Our results showed: 1) at the medium WM load level, the networks in SZ showed a lower clustering coefficient and less local efficiency compared with HC; 2) in SZ, most network measures altered significantly as the WM load level increased from low to medium and from medium to high, while the network metrics were relatively stable in HC at different WM loads; and 3) the altered structure at medium WM load in SZ was related to their performance during the task, with longer reaction time related to lower clustering coefficient and lower local efficiency. These findings suggest brain connectivity in patients with SZ was more diffuse and less strongly linked locally in functional network at intermediate level of WM when compared to HC. SZ show distinctly inefficient and variable network structures in response to WM load increase, comparing to stable highly clustered network topologies in HC.
BibTeX
@inproceedings{he2012altered,
  author = {Hao He and Jing Sui and Qingbao Yu and Jessica A. Turner and Beng‐Choon Ho and Scott R. Sponheim and Dara S. Manoach and Vincent P. Clark and Vince D. Calhoun},
  title = {{Altered Small-World Brain Networks in Schizophrenia Patients during Working Memory Performance}},
  booktitle = {PLoS},
  year = {2012},
  doi = {10.1371/journal.pone.0038195},
}
Figure