PT  - JOURNAL ARTICLE
AU  - Chou, Y.-h.
AU  - Panych, L.P.
AU  - Dickey, C.C.
AU  - Petrella, J.R.
AU  - Chen, N.-k.
TI  - Investigation of Long-Term Reproducibility of Intrinsic Connectivity Network Mapping: A Resting-State fMRI Study
AID  - 10.3174/ajnr.A2894
DP  - 2012 May 01
TA  - American Journal of Neuroradiology
PG  - 833--838
VI  - 33
IP  - 5
4099  - http://www.ajnr.org/content/33/5/833.short
4100  - http://www.ajnr.org/content/33/5/833.full
SO  - Am. J. Neuroradiol.2012 May 01; 33
AB  - BACKGROUND AND PURPOSE: Connectivity mapping based on resting-state fMRI is rapidly developing, and this methodology has great potential for clinical applications. However, before resting-state fMRI can be applied for diagnosis, prognosis, and monitoring treatment for an individual patient with neurologic or psychiatric diseases, it is essential to assess its long-term reproducibility and between-subject variations among healthy individuals. The purpose of the study was to quantify the long-term test-retest reproducibility of ICN measures derived from resting-state fMRI and to assess the between-subject variation of ICN measures across the whole brain. MATERIALS AND METHODS: Longitudinal resting-state fMRI data of 6 healthy volunteers were acquired from 9 scan sessions during >1 year. The within-subject reproducibility and between-subject variation of ICN measures, across the whole brain and major nodes of the DMN, were quantified with the ICC and COV. RESULTS: Our data show that the long-term test-retest reproducibility of ICN measures is outstanding, with >70% of the connectivity networks showing an ICC > 0.60. The COV across 6 healthy volunteers in this sample was >0.2, suggesting significant between-subject variation. CONCLUSIONS: Our data indicate that resting-state ICN measures (eg, the correlation coefficients between fMRI signal-intensity profiles from 2 different brain regions) are potentially suitable as biomarkers for monitoring disease progression and treatment effects in clinical trials and individual patients. Because between-subject variation is significant, it may be difficult to use quantitative ICN measures in their current state as a diagnostic tool. COVcoefficient of varianceDMNdefault mode networkICCintraclass correlation coefficientICNintrinsic connectivity networkIPCinferior parietal cortexITCinferior temporal cortexMPFCmedial prefrontal cortexMTGmiddle temporal gyrusPCCposterior cingulate cortexPHCparahippocampal cortexSFCsuperior frontal cortexVACCventral anterior cingulate cortex