2016/12/13 by Daniel Mirman, Mirman, Daniel, Jon-Frederick Landrigan +9 · 1 citation
Medicine · Neuroscience · #Applications (stat.AP) #Dementia and Cognitive Impairment Research #EEG and Brain-Computer Interfaces #FOS: Biological sciences #FOS: Computer and information sciences #Functional Brain Connectivity Studies #Neural and Behavioral Psychology Studies #Neurons and Cognition (q-bio.NC)
paper · pdf · doi:10.48550/arxiv.1612.04345
openalex publication_date 2016/12/13 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
Voxel-based lesion-symptom mapping (VLSM) is an important method for basic\nand translational human neuroscience research. VLSM leverages modern\nneuroimaging analysis techniques to build on the classic approach of examining\nthe relationship between location of brain damage and cognitive deficits.\nTesting an association between deficit severity and lesion status in each voxel\ninvolves very many individual tests and requires statistical correction for\nmultiple comparisons. Several strategies have been adapted from analysis of\nfunctional neuroimaging data, though VLSM faces a more difficult trade-off\nbetween avoiding false positives and statistical power (missing true effects).\nNon-parametric, permutation-based methods are generally preferable because they\ndo not make assumptions that are likely to be violated by skewed distributions\nof behavioral deficit (symptom) scores and by the necessary spatial contiguity\nof stroke lesions. We used simulated and real deficit scores from a sample of\napproximately 100 individuals with left hemisphere stroke to evaluate two such\npermutation-based approaches. Using permutation to set a minimum cluster size\nidentified a region that systematically extended well beyond the true region,\neven under the most conservative settings tested here, making it ill-suited to\nidentifying brain-behavior relationships. In contrast, generalizing the\nstandard permutation-based family-wise error correction approach provided a\nprincipled way to balance false positives and false negatives. An\nimplementation of this continuous permutation-based FWER correction method is\navailable at https://gist.github.com/dmirman/05a92e0e9e0027f6fe6e528c648143d7\n