Functional connectivity analysis and prediction of cognitive change after carotid artery stenting

Naoki Tani, Takahide Yaegaki, Akio Nishino, Kenta Fujimoto, Hiroyuki Hashimoto, Kaoru Horiuchi, Mitsuhisa Nishiguchi, Haruhiko KishimaJournal of Neurosurgery 2019 · DOI

TL;DR

This study explores how changes in brain network connectivity, specifically the default mode network (DMN), relate to cognitive improvements after carotid artery stenting (CAS). They found that the degree of functional connectivity (FC) between the DMN and areas involved in working memory before CAS can predict how much a patient's working memory will improve after the procedure. This suggests that DMN connectivity could serve as a valuable biomarker for predicting cognitive outcomes in patients undergoing CAS.

What Problem Does This Paper Address?

Cerebral revascularization procedures, such as carotid artery stenting (CAS), aim to prevent strokes in patients with carotid artery stenosis. While the primary goal is stroke prevention, these procedures can sometimes lead to improvements in cognitive function. However, the extent and nature of these cognitive changes vary significantly among patients, and there's a lack of reliable methods to predict who will benefit most cognitively. This variability is a challenge because cognitive impairment is a significant issue, especially in an aging population. The authors hypothesize that alterations in the brain's default mode network (DMN), a network crucial for cognitive processes, might underlie these observed cognitive changes after CAS. They aim to investigate this by examining the relationship between DMN functional connectivity (FC) and post-CAS cognitive function, hoping to identify a predictive marker.

Methods

The study employed a prospective, exploratory design involving 8 male patients with unilateral carotid artery stenosis who underwent CAS.

  • Neuropsychological Testing: Patients underwent a battery of tests, including the Wechsler Adult Intelligence Scale III (WAIS-III) and the Wechsler Memory Scale-Revised (WMS-R), both before and 6 months after CAS. These tests assess various cognitive domains, including working memory, verbal comprehension, and general memory.
  • Resting-State Functional MRI (rs-fMRI): rs-fMRI scans were acquired from each patient before and 6 months after CAS using a 3-T scanner. This technique measures spontaneous brain activity when a person is at rest, allowing researchers to infer functional connectivity between different brain regions.
  • Image Analysis:
    • Preprocessing: Standard fMRI preprocessing steps were applied using SPM12, including realignment, normalization to a standard MNI template, and smoothing.
  • DMN Identification: Independent Component Analysis (ICA) was used to identify the default mode network (DMN) components from the rs-fMRI data. ICA is a data-driven method that separates mixed signals into independent components, allowing for the extraction of specific brain networks. Both anterior (aDMN) and posterior (pDMN) parts of the DMN were identified.
  • Connectivity Analysis: The study analyzed the correlation between changes in DMN functional connectivity (FC) and changes in cognitive function (specifically working memory scores). Voxel-wise correlation analyses were performed.
  • Predictive Modeling: Univariate and multivariate logistic regression analyses were used to identify preoperative factors that could predict post-CAS improvement in working memory.

Datasets

Not discussed in the paper. The study utilized data collected from the 8 participating patients.

Key Insights and Ideas

  • DMN-Working Memory Link: The study strongly suggests that the default mode network (DMN) plays a critical role in cognitive changes observed after carotid artery stenting (CAS), particularly in working memory.
  • Inverse Relationship Post-CAS: Patients who showed improvement in working memory after CAS also exhibited a decrease in functional connectivity (FC) between the posterior DMN (pDMN) and brain regions involved in working memory (precentral/superior frontal gyrus and middle frontal gyrus). This suggests a "normalization" of brain network activity.
  • Predictive Power of Preoperative FC: Crucially, the study found that the preoperative FC between the pDMN and these working memory-related areas was positively correlated with the postoperative improvement in working memory. This is a key finding, as it indicates that a patient's brain network state before the procedure can predict their cognitive outcome.
  • DMN vs. Task-Positive Networks: The DMN is often considered an "anti-correlated" network with task-positive networks (like those involved in attention and working memory). The observed inverse correlation between DMN FC and working memory improvement post-CAS aligns with this concept, suggesting that as task-related networks become more efficient, the DMN's engagement might decrease in specific contexts.

Key Results

  • Cognitive Improvement Varies: As expected, cognitive function changes after CAS varied among patients and across different cognitive tests.
  • Working Memory Changes: While overall working memory scores didn't show a significant average improvement, individual patient improvements were observed.
  • Negative Correlation Post-CAS: Changes in working memory scores after CAS were negatively correlated with changes in FC between the pDMN and the precentral/SFG (p=0.02) and between the pDMN and the MFG (p<0.05).
  • Positive Correlation Pre-CAS: Preoperative FC between the pDMN and the left precentral/SFG correlated significantly with postoperative increases in working memory scores (r=0.93, p=0.0008). Similarly, preoperative FC between the pDMN and the left MFG correlated positively with post-CAS working memory improvement (r=0.81, p=0.01).
  • Predictive Factor: In multivariate analysis, preoperative FC of the pDMN-precentral/SFG emerged as a significant predictor of working memory improvement after CAS.
  • No Stenosis Correlation: Preoperative carotid stenosis severity and changes in stenosis did not correlate with baseline or post-CAS working memory scores, suggesting other factors are more influential for cognitive outcomes in this context.

Limitations and Future Work

The authors acknowledge several limitations:

  • Small Sample Size: The study involved only 8 patients, which limits the statistical power and generalizability of the findings. A power analysis was not performed to determine the optimal sample size.
  • No Brain Perfusion Assessment: The study did not include assessments of brain perfusion, which could provide additional insights into the relationship between blood flow, FC, and cognitive function.
  • Potential Learning Effects: While the 6-month follow-up period was intended to minimize learning effects from repeated testing, it's acknowledged that a minimal learning effect might still be present.
  • Uncertainty on Treatment Side: The small sample size made it difficult to determine if the side of treatment (left vs. right carotid artery) had a significant impact on cognitive outcomes.

Future work suggested includes:

  • Larger, Multicenter Studies: To confirm these preliminary findings and investigate potential effects of treatment side.
  • Inclusion of Brain Perfusion Data: To explore the interplay between cerebral blood flow, FC, and cognition.
  • Longitudinal Investigations: To further understand the mechanisms underlying cognitive decline and restoration, and to track changes over longer periods.
  • Exploring Other Cognitive Domains: While working memory was a focus, exploring other cognitive functions and their relationship with DMN FC would be beneficial.

Summary updated

Abstract

FC between the DMN and working memory-related areas is closely associated with improvement in working memory after CAS. Preoperative analysis of FC of the DMN may be useful for predicting postoperative improvement in the working memory of patients being treated for unilateral stenosis of the extracranial internal carotid artery.Clinical trial registration no.: UMIN000020045 (www.umin.ac.jp/ctr/index.htm).
BibTeX
@article{tani2019functional,
  author = {Naoki Tani and Takahide Yaegaki and Akio Nishino and Kenta Fujimoto and Hiroyuki Hashimoto and Kaoru Horiuchi and Mitsuhisa Nishiguchi and Haruhiko Kishima},
  title = {{Functional connectivity analysis and prediction of cognitive change after carotid artery stenting}},
  journal = {Journal of Neurosurgery},
  year = {2019},
  doi = {10.3171/2018.7.jns18404},
}
Figure