|
Spatial misinference in human brain lesion-function mapping| old_uid | 9396 |
|---|
| title | Spatial misinference in human brain lesion-function mapping |
|---|
| start_date | 2010/12/10 |
|---|
| schedule | 11h-12h30 |
|---|
| online | no |
|---|
| details | invité par A. Gorea |
|---|
| summary | Validating models of the large-scale functional organisation of the brain is critically dependent on the study of patients with focal brain lesions. Since no other comprehensive approach has comparable inferential power - establishing the necessity of a brain region for a putative function - any consistent errors in anatomical localization cannot be readily corrected. Analyzing a very large number of such lesions, here we show that biases in the pattern of damage across the brain critically distort the spatial inferences drawn from the data, making the localisations from all previous studies fundamentally insecure. We quantify the size of the error for the limiting case of a single point critical locus, and show that no reliable estimates can be derived from multi-locus criticality with conventional mass-univariate mapping techniques. The need for a comprehensive re-evaluation of lesion-function data is thereby rendered inevitable. |
|---|
| responsibles | Rämä |
|---|
| |
|