A Neural Model of Recovery from Lesions in the Somatosensory System

Modeling brain reorganization after hemispherectomy

“…The process seems smooth, as units at the edge of the (shrinking) orphaned domain are parsimoniously incorporated. This is consistent with earlier studies of lesions in topographic maps based on SOM [30][31][32]34].…”

“…This hampers the gradual incorporation of orphan stimuli through the map's edge (opposed to what happened in Supporting Figure 7). This mechanism works locally at most, in agreement with earlier simulations of local injuries [30][31][32]34].…”

“…Brain reorganization has been proved, among others, by experiments in macaques [22][23][24][25][26][27], clinical procedures [28], and recovery after stroke [29]. Simple mathematical models [30][31][32][33][34] illustrate smooth, local processes when topographic maps need to expand (invading adjacent areas) or change their nature slightly. Less consensus emerges at larger scales.…”

“…Self-Organized Maps [35] are a class of neural networks able to infer, in an unsupervised fashion, the geometrical structure underlaying a given data set. Different SOM versions have been used to model local plasticity in topographic maps after an injury [30][31][32][33][34]. Our implementation is the simplest possible, faithful to the original scheme.…”

Reemergence of activation with poststroke somatosensory recovery: A serial fMRI case study

“…Evolution of neuroplastic changes in the preexisting somatosensory system is consistent with the theoretical model of cerebral recovery described by Hughlings Jackson 8 and specified for the somatosensory system. 9 In our patient with a subcortical lesion, the change could also be associated with reversible loss of function of thalamic neurons temporarily impaired at the infarct site. Whatever the mechanism, the potential for reemergence of activation in multiple sites typically activated during this task remains.…”