Individualized MRI neuromodulation enhances visual perception targeted towards the rehabilitation of cortical blindness in optic glioma patients
Introduction
The Papageorgiou lab has developed a closed-loop, fMRI-based neuromodulation, we call individualized neuromodulation (iNM) with the long-term goal to rehabilitate visual perception in cortically blind (CB) patients with lesions below the optic chiasm. These lesions result in visual field deficits called hemianopias, which negatively impact patients’ quality of life and productivity. The etiological factors of cortical blindness are multifactorial; the most common one being stroke but also optic nerve glioma tumors. Behavioral training (BT) attempts to restore vision in CB but is a very lengthy, not-targeted, and not an individualized intervention. Our goal was to elucidate the neuromodulation visual perception mechanisms of iRTfMRI0cNMT in healthy participants.
Methods
Fourteen healthy subjects were asked to discriminate up or down motion at 100% or 33% coherence, interleaved with periods of random motion. Study-Day-1 included 10 fMRI scans to localize up and down direction at 100% and 33% motion coherence. Study-Day-2 alternated between 6 iRTfMRI-cNMT and 6 Control-No-cNMT scans.
Results
iNM enhanced the HbO2 magnitude associated with visual motion perception: 1) 48% and 93% increase in the HbO2 magnitude AUC for up and down fully coherent motion and 290% and 316% increase for up and down motion at subthreshold levels, (p<0.005); 2) the variance of the HbO2 magnitude for up motion decreased by ~27% at 100% and ~75% at 33% coherence, and for down motion decreased by 67% and 46%, respectively (p<0.01). Causal modeling identified two brain states associated with visual motion performance across directions and coherences: 1) a dominant state, associated with up motion at 100% coherence in 99% of the iNM trials, while at 33%, it occurred in 76% of the iNM trials, and selectivity for the control trials at 100% coherence occurred in 91% of those, while at 50% of the trials were associated with 33% coherent upward motion (p=0.01); and 2) a non-dominant state identified for random motion when up motion was presented, occurred in 1% of the iNM trials at 100% coherence and 24% at 33% coherence, while 7% of the control trials were identified as random motion when 100% coherence was presented and 50% of the trials at 33% coherence (p=0.01); 3) a dominant state, during down motion at 100% coherence selected cortical areas associated with down motion 100% of the iNM trials, while at 33% it selected for 96% of the iNM trials, and selectivity for the control trials at 100% coherence occurred in 95% of those, while 96% of the trials were associated with 33% coherent upward motion (p=0.01); and 4) a non-dominant state for random motion during the presentation of down motion occurred in 0% of the iNM trials at 100% coherence and 4% of iNM trials at 33% coherence, while 5% of the control trials were identified as random motion when 100% coherence was presented and 10% of the trials at 33% coherence (p=0.01). Additionally, we decoded the brain states as a function of direction and coherence and found that iNM increased classification accuracies compared to the control-no iNM condition under both coherences: 1) at 100% coherence for the up (iNM: 79%) and down (iNM: 81%) direction discriminations versus 67% and 66%, respectively; and 2) at 33% coherence for the up (iNM: 73%) and down (iNM: 73%) directions versus 68% and 67%.
Conclusion
This study shows spatiotemporal causality between the cortical states engaged during motion perception and the networks involved via iNM. Modularity of sensory (cerebellum; ACC; cingulate), executive and visual motor (orbitofrontal; middle, and superior frontal) and motion perception (intraparietal lobule, precuneus) networks via iNM can serve as biomarkers in the rehabilitation of motion perception in cortically blind optic glioma patients.