Poster K7 - CROSSMODAL CONGURENCY EFFECT AS A MEASURE FOR SENSORY FEEDBACK IN LOWER LIMB AMPUTEES

Thursday, October 24, 2024

2:45 PM - 3:45 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

CO

Charles A. Oduro, Jr. (he/him/his)

Undergraduate Student Researcher
University of Pittsburgh/Rehab Neural Engineering Labs
Pittsburgh, Pennsylvania, United States

Co-Author(s)

RB

Rohit Bose

Mentor/Trainer
University of Pittsburgh/Rehab Neural Engineering Labs, United States

Primary Investigator(s)

LF

Lee Fisher, PhD

Primary Investigator
University of Pittsburgh/Rehab Neural Engineering Labs, Pennsylvania, United States

Introduction: The loss of sensory feedback following lower-limb amputation disrupts balance and gait, making its restoration essential. Methods like multiple invasive and noninvasive stimulation strategies can be used to deliver the sensory feedback. Sensory integration is the process by which the brain combines information from different sensory modalities (e.g., visual, auditory, tactile) to form a coherent perception and response. Effective sensory integration is crucial for smooth and coordinated motor functions, especially in complex tasks that require precise movements and adjustments based on sensory feedback. There are multiple strategies to provide sensory feedback, the challenge being determining which strategy can effectively integrate into the user's existing sensorimotor neural schema (the brain's internal representation of the body and its movements). If the sensory feedback does not integrate well, it may not be useful or could even be confusing for the user. To assess sensory integration, we employed the cross-modal congruency task. This task involves simultaneous delivery of two sensory modalities to different body locations. Participants report the location of a target stimulus while disregarding a distractor stimulus. The crossmodal congruency effect (CCE) measures the difference in brain activity between the two sensory modality locations. This effect quantifies the integration between different sensory systems in the brain. CCE was recently proposed as a measure of sensory feedback integration in upper limb amputees [1]. This study aims to find cortical biomarkers of the crossmodal congruency effect better understand the underlying neural processes of sensory integration during restored sensory feedback mechanisms for lower limb prosthetics.

Materials and

Methods: Twelve able-bodied participants, aged 18 to 65, with normal or corrected-to-normal vision, participated in the study. Participants sat with their right leg extended, where LEDs, electrical stimulation electrodes, and pneumatic tactors were placed on the foot and the knee, with an additional LED on the shin serving as a visual fixation point. Simultaneous somatosensory and visual stimuli were delivered either in the same location (congruent trials) or a different location (incongruent trials). Participants reported where they felt the somatosensory stimulus while ignoring the visual stimulus. The experiment included six blocks, each with 12 congruent and 12 incongruent trials for each location and stimulus type (pneumatic or electric), with 5-10 minute breaks between blocks. EEG data were recorded throughout and processed in EEGLAB, where they were down sampled to 500 Hz, filtered using a FIR filter with a [0.3 40] Hz bandwidth, and epochs were extracted from [-2 to 2.5] seconds around stimulus onset, later being used for time-locked analysis. Independent component analysis was used to remove artifacts through EEGLAB. Extracted epochs were normalized using the baseline correction technique, reducing variability. The data was analyzed in both stimulus-locked [0 200] milliseconds and response-locked [-200 0] milliseconds windows to allow for targeted analysis of distinct neural processes associated with sensory stimuli and motor responses, respectively [2]. This combined approach helps us understand how the brain reacts to sensory inputs and how it controls movements at different times. The data was averaged by condition and location and analyzed using a Wilcoxon rank-sum test.

Results, Conclusions, and Discussions: Figure 1 displays the EEG topographic maps of the stimulus-locked normalized EEG data of the difference between the incongruent and congruent trials. An opposite effect was observed between the knee and foot trials during the pneumatic and electric conditions. The topographic maps reveal differential neural processing between congruent and incongruent trials for knee and foot stimuli. In the parietal electrodes, the knee trials show a higher amplitude for congruent conditions, indicating stronger neural activity during congruent trials compared to incongruent ones. Conversely, the foot trials show a higher amplitude for incongruent conditions, suggesting that the neural activity is more pronounced during incongruent trials compared to congruent ones. These findings suggest that the parietal region differentially processes congruency depending on whether the stimulus involves the knee or the foot, with greater engagement for congruent stimuli in knee trials and for incongruent stimuli in foot trials. Figure 2 displays the results of the Wilcoxon rank-sum test of the stimulus-locked normalized EEG data of the difference between the incongruent and congruent trials. The analysis revealed that only three to four participants exhibited statistically significant differences between the incongruent and congruent conditions. This limited number of participants showing significance indicates that CCE on the EEG signals was not consistently observed across the sample. The lack of widespread significant findings across participants suggests that EEG-based biomarkers may not be sufficiently effective in differentiating between the crossmodal congruency effects under the conditions tested.

Using time-locking analysis, there is an opposite effect occurring in the parietal electrodes between the knee and foot trials. This effect was observed in the delta and theta bands, but with few subjects exhibiting statistical significance, we determined that EEG-based analysis of CCE is not a good measure for sensory integration. Further research can be done to build on these findings such as including lower limb amputees as participants with the goal to understand how amputees perceive and integrate sensory feedback. Additionally, exploration of different modulates and/or stimuli to understand the robustness and specificity of the effect across different sensory modalities can be studied to help tailor sensory feedback in prosthetics.

Acknowledgements (Optional):