Microstimulation for Somatosensory Neuroprosthesis: Mapping direct responses and afferent modulation in somatosensory cortex and ventral posterolateral thalamus subnuclei.

Abstract

Brain machine interfaces (BMIs) aim to restore lost motor functions by channeling movement-related brain signals to end-effectors skipping compromized parts of the central nervous system (CNS.) As there is increasing number of patients who can potentialy benefit from such help, there is a very urgent demand for optimizing and controlling the outputs of such devices. These BMI devices are currently is far from performing the every day and seemingly unintelligent acts of reaching, grasping, object manipulation, force-adaptation and motor learning. This is because these acts call on the CNS to make complex computations on its visual, proprioceptive and cutaneous inputs which artificial limbs lack, the main reason why even the best BMIs still give unstable movements. Efforts are now underway to create artificial proprioceptive and cutaneous inputs to BMIs. However, due to lack of thorough understanding of its effects on the brain and BMI movements, the presently accepted candidate for such a feedback channel, Microstimulation (MiSt) of the primary somatosensory cortex (S1), is not yet fully functional. Numerous critical questions have yet to be addressed, such as which S1 area and layer to target, what stimulus parameters to use, if unwanted side-effects, such as brain lesions, movement etc. should be major concerns, and many more. Moreover, we asked, why not even microstimulate other brain areas? Currently, there is little information on MiSt for somatosensation in other brain areas, such as the thalamus. This thesis aimed at: 1) Finding the discernable differences and similarities between the two inputs, natural and artificial (MiSt in S1 cortex and the VPL thalamus); 2) Finding the best region, cortex or VPL, where MiSt gives the most natural-looking response in S1 cortex; 3) Studying the dynamics of causal information transfer in thalamocortical networks in natural and MiSt for somatosensory neuroprosthesis to investigate possible behavioral consequences. MiSt was applied to S1 cortex and VPL thalamus of a monkey (multiple implants) and rats implanted with multielectrode and multiwire arrays respectively. The resulting spike and Local Field Potential (LFP) cortical and thalamic activities were recorded and analyzed per the goals of the thesis. The main findings indicate that cortical responses in both input conditions are comparable and that VPL MiSt may be a better candidate for somatosensory neuroprosthesis mainly because it causes reduced or no cortical inhibition. The results show that the different stimulation conditions have specific effects on the causal interaction in and between the two brain regions.