Measurement of sensory deficiency in fine touch after stroke during textile fabric stimulation by electroencephalography (EEG)

Somatosensory integration in robot-assisted motor restoration post-stroke

Disruption of somatosensorimotor integration (SMI) after stroke is a significant obstacle to achieving precise motor restoration. Integrating somatosensory input into motor relearning to reconstruct SMI is critical during stroke rehabilitation. However, current robotic approaches focus primarily on precise control of repetitive movements and rarely effectively engage and modulate somatosensory responses, which impedes motor rehabilitation that relies on SMI. This article discusses how to effectively regulate somatosensory feedback from target muscles through peripheral and central neuromodulatory stimulations based on quantitatively measured somatosensory responses in real time during robot-assisted rehabilitation after stroke. Further development of standardized recording protocols and diagnostic databases of quantitative neuroimaging features in response to post-stroke somatosensory stimulations for real-time precise detection, and optimized combinations of peripheral somatosensory stimulations with robot assistance and central nervous neuromodulation are needed to enhance the recruitment of targeted ascending neuromuscular pathways in robot-assisted training, aiming to achieve precise muscle control and integrated somatosensorimotor functions, thereby improving long-term neurorehabilitation after stroke.

Comparison of Immediate Neuromodulatory Effects between Focal Vibratory and Electrical Sensory Stimulations after Stroke

Focal vibratory stimulation (FVS) and neuromuscular electrical stimulation (NMES) are promising technologies for sensory rehabilitation after stroke. However, the differences between these techniques in immediate neuromodulatory effects on the poststroke cortex are not yet fully understood. In this research, cortical responses in persons with chronic stroke (n = 15) and unimpaired controls (n = 15) were measured by whole-brain electroencephalography (EEG) when FVS and NMES at different intensities were applied transcutaneously to the forearm muscles. Both FVS and sensory-level NMES induced alpha and beta oscillations in the sensorimotor cortex after stroke, significantly exceeding baseline levels (p < 0.05). These oscillations exhibited bilateral sensory deficiency, early adaptation, and contralesional compensation compared to the control group. FVS resulted in a significantly faster P300 response (p < 0.05) and higher theta oscillation (p < 0.05) compared to NMES. The beta desynchronization over the contralesional frontal-parietal area remained during NMES (p > 0.05), but it was significantly weakened during FVS (p < 0.05) after stroke. The results indicated that both FVS and NMES effectively activated the sensorimotor cortex after stroke. However, FVS was particularly effective in eliciting transient involuntary attention, while NMES primarily fostered the cortical responses of the targeted muscles in the contralesional motor cortex.

Identifying Oscillations under Multi-site Sensory Stimulation for High-level Peripheral Nerve Injured Patients：A Pilot Study

OBJECTIVE

For high-level peripheral nerve injured (PNI) patients with severe sensory dysfunction of upper extremities, identifying the multi-site tactile stimulation is of great importance to provide neurorehabilitation with sensory feedback. In this pilot study, we showed the feasibility of identifying multi-site and multi-intensity tactile stimulation in terms of electroencephalography (EEG).

APPROACH

Three high-level PNI patients and eight non-PNI participants were recruited in this study. Four different sites over the upper arm, forearm, thumb finger and little finger were randomly stimulated at two intensities (both sensory-level) based on the transcutaneous electrical nerve stimulation (TENS). Meanwhile, 64-channel EEG signals were recorded during the passive tactile sense stimulation on each side.

MAIN RESULTS

The spatial-spectral distribution of brain oscillations underlying multi-site sensory stimulation showed dominant power attenuation over the somatosensory and prefrontal cortices in both alpha-band (8-12 Hz) and beta-band (13-30 Hz). But there was no significant difference among different stimulation sites in terms of the averaged power spectral density over the region of interest (ROI). By further identifying different stimulation sites using temporal-spectral features, we found the classification accuracies were all above 89% for the affected arm of PNI patients, comparable to that from their intact side and that from the non-PNI group. When the stimulation site-intensity combinations were treated as eight separate classes, the classification accuracies were ranging from 88.89% to 99.30% for the affected side of PNI subjects, similar to that from their non-affected side and that from the non-PNI group. Other performance metrics, including Specificity, Precision, and F1-Score, also showed a sound identification performance for both PNI patients and non-PNI subjects.

SIGNIFICANCE

These results suggest that reliable brain oscillations could be evoked and identified well, even though induced tactile sense could not be discerned by the PNI patients. This study have implication for facilitating bidirectional neurorehabilitation systems with sensory feedback.

Electroencephalographic Measurement on Post-stroke Sensory Deficiency in Response to Non-painful Cold Stimulation

Background

Reduced elementary somatosensation is common after stroke. However, the measurement of elementary sensation is frequently overlooked in traditional clinical assessments, and has not been evaluated objectively at the cortical level. This study designed a new configuration for the measurement of post-stroke elementary thermal sensation by non-painful cold stimulation (NPCS). The post-stroke cortical responses were then investigated during elementary NPCS on sensory deficiency via electroencephalography (EEG) when compared with unimpaired persons.

Method

Twelve individuals with chronic stroke and fifteen unimpaired controls were recruited. A 64-channel EEG system was used to investigate the post-stroke cortical responses objectively during the NPCS. A subjective questionnaire of cold sensory intensity was also administered via a numeric visual analog scale (VAS). Three water samples with different temperatures (i.e., 25, 10, and 0°C) were applied to the skin surface of the ventral forearm for 3 s via glass beaker, with a randomized sequence on either the left or right forearm of a participant. EEG relative spectral power (RSP) and topography were used to evaluate the neural responses toward NPCS with respect to the independent factors of stimulation side and temperature.

Results

For unimpaired controls, NPCS initiated significant RSP variations, mainly located in the theta band with the highest discriminative resolution on the different temperatures (P < 0.001). For stroke participants, the distribution of significant RSP spread across all EEG frequency bands and the temperature discrimination was lower than that observed in unimpaired participants (P < 0.05). EEG topography showed that the NPCS could activate extensive and bilateral sensory cortical areas after stroke. Significant group differences on RSP intensities were obtained in each EEG band (P < 0.05). Meanwhile, significant asymmetry cortical responses in RSP toward different upper limbs were observed during the NPCS in both unimpaired controls and participants with stroke (P < 0.05). No difference was found between the groups in the VAS ratings of the different temperatures (P > 0.05).

Conclusion

The post-stroke cortical responses during NPCS on sensory deficiency were characterized by the wide distribution of representative RSP bands, lowered resolution toward different temperatures, and extensive activated sensory cortical areas.

Evaluation of Post-Stroke Impairment in Fine Tactile Sensation by Electroencephalography (EEG)-Based Machine Learning

Electroencephalography (EEG)-based measurements of fine tactile sensation produce large amounts of data, with high costs for manual evaluation. In this study, an EEG-based machine-learning (ML) model with support vector machine (SVM) was established to automatically evaluate post-stroke impairments in fine tactile sensation. Stroke survivors (n = 12, stroke group) and unimpaired participants (n = 15, control group) received stimulations with cotton, nylon, and wool fabrics to the different upper limbs of a stroke participant and the dominant side of the control. The average and maximal values of relative spectral power (RSP) of EEG in the stimulations were used as the inputs to the SVM-ML model, which was first optimized for classification accuracies for different limb sides through hyperparameter selection (γ, C) in radial basis function (RBF) kernel and cross-validation during cotton stimulation. Model generalization was investigated by comparing accuracies during stimulations with different fabrics to different limbs. The highest accuracies were achieved with (γ = 21, C = 23) for the RBF kernel (76.8%) and six-fold cross-validation (75.4%), respectively, in the gamma band for cotton stimulation; these were selected as optimal parameters for the SVM-ML model. In model generalization, significant differences in the post-stroke fabric stimulation accuracies were shifted to higher (beta/gamma) bands. The EEG-based SVM-ML model generated results similar to manual evaluation of cortical responses to fabric stimulations; this may aid automatic assessments of post-stroke fine tactile sensations.

Pathway-specific cortico-muscular coherence in proximal-to-distal compensation during fine motor control of finger extension after stroke

Objective.Proximal-to-distal compensation is commonly observed in the upper extremity (UE) after a stroke, mainly due to the impaired fine motor control in hand joints. However, little is known about its related neural reorganization. This study investigated the pathway-specific corticomuscular interaction in proximal-to-distal UE compensation during fine motor control of finger extension post-stroke by directed corticomuscular coherence (dCMC).Approach.We recruited 14 chronic stroke participants and 11 unimpaired controls. Electroencephalogram (EEG) from the sensorimotor area was concurrently recorded with electromyography (EMG) from extensor digitorum (ED), flexor digitorum (FD), triceps brachii (TRI) and biceps brachii (BIC) muscles in both sides of the stroke participants and in the dominant (right) side of the controls during the unilateral isometric finger extension at 20% maximal voluntary contractions. The dCMC was analyzed in descending (EEG → EMG) and ascending pathways (EMG → EEG) via the directed coherence. It was also analyzed in stable (segments with higher EMG stability) and less-stable periods (segments with lower EMG stability) subdivided from the whole movement period to investigate the fine motor control. Finally, the corticomuscular conduction time was estimated by dCMC phase delay.Main results.The affected limb had significantly lower descending dCMC in distal UE (ED and FD) than BIC (P< 0.05). It showed the descending dominance (significantly higher descending dCMC than the ascending,P< 0.05) in proximal UE (BIC and TRI) rather than the distal UE as in the controls. In the less-stable period, the affected limb had significantly lower EMG stability but higher ascending dCMC (P< 0.05) in distal UE than the controls. Furthermore, significantly prolonged descending conduction time (∼38.8 ms) was found in ED in the affected limb than the unaffected (∼26.94 ms) and control limbs (∼25.74 ms) (P< 0.05).Significance.The proximal-to-distal UE compensation in fine motor control post-stroke exhibited altered descending dominance from the distal to proximal UE, increased ascending feedbacks from the distal UE for fine motor control, and prolonged descending conduction time in the agonist muscle.

Neuro-perceptive discrimination on fabric tactile stimulation by Electroencephalographic (EEG) spectra

The precise evaluation of sensory perceptions during fabric-skin interactions is still poorly understood in neuroscience. This study aims to investigate the cortical sensory response to fabric stimuli with different textiles by Electroencephalographic (EEG) spectral intensities, and evaluate the relationships between EEG frequency bands, traditional subjective questionnaires, and the materials' physical properties. Twelve healthy adult participants were recruited to test three fabrics with different textile compositions of 1) cotton, 2) nylon, and 3) polyester and wool. The physical properties of the fabrics were quantitatively evaluated by a Fabric Touch Tester (FTT). Subjects were invited to rate the sensory perception of the fabric samples via a subjective questionnaire and objective EEG recording. Significant differences in the EEG relative spectral power of Theta and Gamma bands were acquired in response to the different fabric stimuli (P<0.05). The Theta and Gamma powers demonstrated a significant correlation with the most of the subjective sensations evaluated by questionnaire and the fabrics' physical properties by FTT (P<0.05). The EEG spectral analysis could feasibly be used for the discrimination of fabric stimuli with different textile compositions and further indicates sensory perceptions during fabric stimulation. This finding may provide evidence for further exploratory research of sensory perceptions via EEG spectral analysis, which could be applied to the study of brain generators of skin tactility in future prostheses and the automatic detection of sensory perception in industries.