Journal of the International Neuropsychological Society

Symposia

Multivariate Pattern Analysis of fMRI in Breast Cancer Survivors and Healthy Women

S.M. Hadi Hosseinia1 and Shelli R. Keslera1a2 c1

a1 Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California

a2 Stanford Cancer Institute, Palo Alto, California

Abstract

Advances in breast cancer (BC) treatments have resulted in significantly improved survival rates. However, BC chemotherapy is often associated with several side effects including cognitive dysfunction. We applied multivariate pattern analysis (MVPA) to functional magnetic resonance imaging (fMRI) to find a brain connectivity pattern that accurately and automatically distinguishes chemotherapy-treated (C+) from non-chemotherapy treated (C−) BC females and healthy female controls (HC). Twenty-seven C+, 29 C−, and 30 HC underwent fMRI during an executive-prefrontal task (Go/Nogo). The pattern of functional connectivity associated with this task discriminated with significant accuracy between C+ and HC groups (72%, p = .006) and between C+ and C− groups (71%, p = .012). However, the accuracy of discrimination between C− and HC was not significant (51%, p = .46). Compared with HC, behavioral performance of the C+ and C− groups during the task was intact. However, the C+ group demonstrated altered functional connectivity in the right frontoparietal and left supplementary motor area networks compared to HC, and in the right middle frontal and left superior frontal gyri networks, compared to C−. Our results provide further evidence that executive function performance may be preserved in some chemotherapy-treated BC survivors through recruitment of additional neural connections. (JINS, 2013, 19, 1–11)

(Received March 01 2013)

(Revised September 24 2013)

(Accepted September 25 2013)

(Online publication October 18 2013)

Keywords

  • Machine learning;
  • Response inhibition;
  • fMRI;
  • Functional connectivity;
  • Breast cancer;
  • Executive function

Correspondence

c1 Correspondence and reprint requests to: Shelli Kesler,401 Quarry Road, MC5795,Stanford, CA 94305-5795. E-mail: skesler@stanford.edu

Related Content