ABSTRACT
Objectives: Functioning, but injured cerebral connections are hypothesized to inhibit cortical plasticity. Study of neural networks can validate this hypothesis, and provide further practical clues for clinical and surgical options to restore function in eloquent brain areas.
Material and methods: Cortical lesions in eloquent areas were simulated by means of artificial neural networks. Next, functional restoration of these networks after lesional bypass was studied.
Results: The accuracy of network outputs was reduced from 92% to 72% (P-value < 0.001) when logical temporal connections with dysfunctional lesions were established. Restoration of function was almost totally achieved by bypassing the lesion, without any significant changes in network nodal weights. Estimated remaining functional fraction errors were trivial (0.0044%–1.4%).
Discussion: Examples of functional decline due to disturbing signals are Todd's paralysis and Landau–Kleffner syndrome. Functional restoration after lesionectomy in eloquent areas of the brain is also practiced. Likewise, injured connections provide routes of influence for disturbing impulses.
Conclusion: Herein, the proposed evidences provide theoretical clues to formulate new avenues in restorative functional neurosurgery. They may help to identify suitable lesions and suitable techniques for functional restoration including dissection of disturbing connections, bridging and bypassing lesions that can be corroborated by simulation.
Acknowledgments
Special thanks to Mr Mojtaba Neydavoudi for his assistance in statistical analysis and to Mr H. Argasi at the Research Consultation Center (RCC) affiliated to Shiraz University of Medical Sciences for his assistance in language edition.
Disclosure statement
No potential conflict of interest was reported by the author.