Format

Send to

Choose Destination
Neuroimage. 2008 Jul 1;41(3):970-84. doi: 10.1016/j.neuroimage.2007.12.033. Epub 2007 Dec 27.

Thresholding lesion overlap difference maps: application to category-related naming and recognition deficits.

Author information

1
Laboratory of Computational Neuroimaging, Department of Neurology, Division of Behavioral Neurology and Cognitive Neuroscience, University of Iowa College of Medicine, 200 Hawkins Drive, Iowa City, IA 52242, USA.

Abstract

Lesion overlap difference maps have been used in studies designed to test anatomical hypotheses regarding brain systems critical for various cognitive and behavioral tasks, including naming and recognition of concrete entities [Damasio, H., Tranel, D., Grabowski, T., Adolphs, R., Damasio, A., 2004. Neural systems behind word and concept retrieval. Cognition 92, 179-229]. To date, the interpretation of these results has focused on areas of maximum lesion overlap differences. Here we explore formal methods for statistical thresholding and power analysis. We derive exact voxel-wise statistics describing the behavior of lesion overlap difference maps and lesion proportion difference maps under the null hypothesis of no association between lesion and deficit, and we apply the statistics to a large subset of the subjects previously reported in [Damasio, H., Tranel, D., Grabowski, T., Adolphs, R., Damasio, A., 2004. Neural systems behind word and concept retrieval. Cognition 92, 179-229], in order to reassess the lesion correlates of deficits in naming and recognition for five categories of concrete entities. The thresholded maps confirmed many of the results reported previously, but also revealed some differences. Differences in spatial distribution of the lesion correlates of impaired naming of unique versus nonunique entities were confirmed in the inferotemporal region (IT), although overlapping components across categories became apparent in left IT. Additionally, the left inferior frontal gyrus (IFG) was implicated in naming both categories of nonunique natural entities (animals and fruits/vegetables). In corresponding power analyses, we estimated where significant effects could be found under an assumption of maximal effect size given the observed spatial distribution of lesions. Such "effective coverage maps" are valuable for the interpretation of the results, notably because of heterogeneity in lesion coverage encountered in lesion studies. We strongly suggest that when inferential statistics are used in voxel-wise lesion-deficit statistical mapping, these or other power maps be included in the reports.

PMID:
18442925
PMCID:
PMC2582202
DOI:
10.1016/j.neuroimage.2007.12.033
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Elsevier Science Icon for PubMed Central
Loading ...
Support Center