While walking is an activity most people do every day, how the brain functions during this movement remains an open scientific question. The project goal is to determine what brain areas are active, what roles they play, and how they interact throughout specific walking and walking-like behaviors. To accomplish this, electrocortical simulations will be used to determine what specific aspects of interaction are captured by a variety of functional connectivity measures. Next, independent component analysis will identify neurological sources from EEG data recorded during standing and walking in healthy and Parkinson's subjects, and in healthy subjects performing a simplified version of walking. Finally, functional connectivity will be measured between EEG signals and electrocortical sources in brain regions such as the anterior cingulate and the left and right sensorimotor cortices during these conditions to determine how the brain functions move the body in a changing environment.
This research will provide the fellow with experience in computational neuroscience. The project's results may benefit compromised populations by improving rehabilitation practices, such as in analysis and treatment of movement disorders. Understanding brain function in individuals with Parkinson's or spinal cord injury, could potentially lead to assistive devices that restore motor function. Further, this project will aim to increase the number of women involved in science by mentoring undergraduates and by putting on a camp for young girls that uses hands-on labs to show how neuroscience and math can explain simple everyday movements.
I originally planned to use electroencephalography (EEG) in combination with mathematical analyses to examine how different areas of the brain interact during walking and similar activities. I then wanted to determine how this interaction differs in individuals with Parkinson's. However, while analyzing walking EEG data, I discovered that movement during walking induced changes in the sensors that measure brain activity. The effect of movement cannot be separated from the brain data after collection using current methods. I needed to solve this problem to do further study. My colleagues and I devised a way to use EEG sensors to measure just the changes in the sensors due to movement, the "movement artifact." We had people wear a non-conductive cap to block any electrical signals from their bodies and collected EEG on a simulated scalp on top of this (Figure 1). We found that the movement artifact showed specific patterns that are linked to the walking gait cycle. These patterns varied across sensor location, subject, and speed (Figure 2). In particular, there was little artifact at slower speeds. However, when averaged, movement artifact patterns at faster speeds look similar to previously published data on EEG during walking, suggesting that previous data was contaminated by movement artifact. We then tested the effect of movement artifact on mathematical analyses for brain data by using these techniques to process pure movement artifact. We first performed independent component analysis (ICA) and dipole fitting, which are used to identify and locate neural and other sources. Most sources had locations outside the brain, but 1% of sources had locations inside the brain with basic neural characteristics. However, further analysis of these sources showed that these sources lacked other neural characteristics and likely result from ICA combining multiple sources outside the brain. These results show the limitations of using ICA on EEG data and can advise researchers on interpretation of sources that may include movement artifact. We also tracked how connectivity, measurement of how different brain areas interact, between EEG channels was affected by movement artifact. For all measures tested, movement artifact caused connectivity values to fluctuate over a walking stride. However, these fluctuations were specific to subject, speed, and pairs of channels. Further, some measures had smaller connectivity magnitudes, showing that they handled movement artifact better than other measures. However, because all measures showed some fluctuations, we may need more sophisticated measures of connectivity. I have explored possibilities and have found a potential measure, but will need to adjust it slightly for use with walking data. The connectivity results reveal the limitations of current connectivity measures and suggest changes for the future. Lastly, we analyzed whether we could obtain good brain data for neural tasks performed during walking. We found that we could wash out the walking linked patterns in movement artifact data by time-locking this data to a non walking-related event and averaging across trials. We then compared connectivity values from movement artifact data to those from neural data from people performing a cognitive task while both walking and standing. There were similar patterns of brain interaction during walking and standing. For some measures, connectivity magnitudes for these conditions were much larger than those that were present in movement artifact (Figure 3). We therefore showed that it is possible to get good connectivity data for other tasks performed during walking. The results of these studies have ramifications for neuroscience and beyond. Characterizing artifact can help us reinterpret previous results and mitigate problems in the future, allowing for tracking of EEG during everyday activities. Understanding which measures of connectivity can track changes in data with artifact can help signal analysis across fields. Proving we can track brain interaction in cognitive tasks during walking opens up new possibilities in cognitive research. This work can help us track neural changes during everyday tasks, with applications to brain machine interface research, and in individuals with brain disorders, with applications for rehabilitation. In addition to my research, the past two summers, I created, designed curriculum for, and put on Girls With Nerve, a neuroscience camp for high school girls (Figure 4). The goal of the camp is to empower girls to explore neuroscience, and by doing so increase female involvement in STEM fields. The camp is free to participants and runs for a week each summer. My students do hands-on experiments and have discussions where they get to explore neuroscience ideas. They also visit neuroscience research labs so that they can see cutting-edge work, and meet female neuroscientists so that they can learn about what career opportunities they might have in neuroscience. So far, my students have both enjoyed and learned a lot from the camp, and I hope to continue it in the future. A link to the webpage can be found here: https://sites.google.com/a/umich.edu/girls-with-nerve/home
