NSF EPSCoR: RII Track-2 FEC: Innovative, Broadly Accessible Tools for Brain Imaging, Decoding, and Modulation

Grants and Contracts Details


Specific Aims for University of Kentucky sub-award Specific Aim 1: Interactive modulation of sensorimotor activity using a brain-machine interface (Sunderam). Rationale: BMIs are increasingly used not just as assistive devices for individuals with motor impairment but to augment rehabilitative treatment and to interact with media such as video games. The more we use such devices, the more our neuroplastic brains adapt to their design in subliminal ways. Hence it is useful to know how the sensorimotor rhythm of the brain is or can be modified by BMI use and whether the mode of interaction can be tuned/designed to optimize the potential benefit and efficacy of these devices. Task 1 (Year 1): Algorithms for decomposition of movement-related EEG dynamics. Sensorimotor BMI commonly relies on imagery tasks in which the exact timing of specific events such as user volition and task compliance are not directly observable or verifiable (e.g., visual evoked response to a cue, motor planning, movement) and must be inferred with reference to sensory cues or detected changes in the EEG. We have developed a computational framework based on hidden Markov models (HMMs), in which the dynamics of the latent states and transitions between them can be inferred from the EEG using maximum likelihood techniques (Fig. 1). Our approach is novel in that a higher order model can be used to obtain a more fine-grained decomposition than the human-recognized brain states. We will use our modeling approach to further decompose combined EEG-fNIRS recordings into movement-related dynamics associated with cue recognition, motor planning, initiation, and control of exerted force in an interactive movement task. The ability to segment the fine structure of EEG associated with imagined and actual movements will be investigated. These developments will increase the degrees of freedom available to the interface and facilitate accurate BMI control in subjects with movement disabilities. Task 2 (Years 2-3): Feedback modulation of the sensorimotor rhythm using an interactive motor task. The sensorimotor or "mu" rhythm serves as the basis for many EEG BMIs, but the manifestation of the mu rhythm in the individual can be quite variable. Using the algorithms developed above to decode EEG-fNIRS signals associated with cued interactions in real time, we will perform a study to test whether the strength of the mu rhythm correlates with motor effort. A pilot study in our lab, in which subjects were prompted to control hand force at random levels in response to a cue, suggested that force prediction from the EEG may be feasible using generalized linear models. Once we have a model that correlates mu rhythm strength with motor effort in baseline trials, we will provide interactive feedback to the user in the form of predicted rather than actual force to modulate their mu rhythm using motor imagery (imagined movement). Task 3 (Years 3-4): Comparison of off-the-shelf equipment with URI tEEG-fNIRS system. The final task will be a comparison of performance on Tasks 1-3 of the novel EEG-fNIRS system being developed by the URI team with off-the-shelf hardware for EEG and fNIRS imaging; for the fNIRS, we will get assistance from Dr. Guoqiang Yu's lab. Metrics compared will be the strength of correlation between EEG features and exerted force in a hand grip task, and user performance on the force control task using motor imagery. It is expected that the TCRE system developed at URI will provide us with a superior ability to localize and discriminate task-related activity compared to the differential and large Laplacian derivations of the 10-20 montage used by most investigators. Figure 1. HMM classifier output for a sample overnight sleep recording. Input features S1, S2, and S3 are shown below the model-generated (black) and true (beige) hypnograms for the data (Cohen’s kappa = 0.8). Specific Aim 2: Entrainment of heart rhythms by auditory evoked activity (Patwardhan): Rationale: It has long been recognized that certain auditory stimuli can have a pronounced effect on cardiac and respiratory rhythms. For example, many young individuals react quite strongly when exposed to certain types of music; the reaction sometimes includes the feeling that their heart rates become synchronous with what is perceived as the dominant rhythm of the music. In many cultures, men and women employ rhythmic music to enter trance-like states with individuals often exhibiting syncopal or pre-syncopal physiological responses. Anecdotal reports of effects of music, particularly rhythmic music, on function of the heart and breathing abound, increasingly, controlled studies in the laboratory and hospital settings do support these reports. Recently there has been an increasing recognition of the palliative effects of music on recovery in a hospital setting and during pre-operative stage. A recent study even suggests that the anxiety reducing effect of music in a preoperative setting may be comparable to that of midazolam. While these results provide a holistic view of the effects of music, mechanistic studies show that rhythmic components in music do affect cardiovascular and cerebrovascular regulation. The current understanding of the mechanisms via which rhythmic auditory stimuli modulate the cardiovascular and cerebrovascular rhythms is not complete. SA 2.a: To quantify, in young adults, entrainment between rhythms in neural, cardiovascular, and respiratory patterns as a response to rhythmic auditory stimuli. We will use two types of auditory stimuli, one that is perceived to have a fast rhythm (tempo) and one that has a slow rhythm. Silence will serve as control. All subjects will be presented with the same stimuli. The sequence of the stimuli will be randomized among subjects. SA 2.b: To determine whether cortical involvement impacts the crossover and entrainment that occurs between rhythmic auditory stimuli and autonomic rhythms. For these experiments, subjects will be presented with two sets of stimuli similar to those for 6.a except that in the first set the language of the songs containing the fast and slow rhythms will be that which is un-intelligible to the subjects. In the second set, the auditory samples will be scrambled in terms of phase, i.e. while the amplitude spectrum will be the same, the phase will be randomized. Specific Aim 3: Effect of meditation on psychomotor vigilance and cognitive function (O'Hara and Sunderam): Rationale: In a previous study conducted by our group to study the effect of meditation on performance, a significant short-term performance boost was seen following meditation in both novice and experienced meditators. The objective of the current project is to analyze effects of meditation on long term performance using physiological measurements of brain dynamics in conjunction with cognitive testing. SA 3.a: To study the effect of meditation on long term performance. Meditators and non-meditators will be evaluated by monitoring psychomotor vigilance task (PVT) scores through the day. The hypothesis is that the performance boost achieved after meditation may be affected by dominance of alpha waves (alpha-power) in the cortex, which may allow the brain time to reset or restore optimal function. Electroencephalography (EEG) will be used to study the brain activity in both experienced and novice meditators. It is predicted that the temporal effect on attention and performance will exhibit a correlation with intensity of alpha waves (alpha power) in the EEG during meditation. In addition we will examine the effect on heart rate, blood pressure and related variables to explore the heart-brain nexus. SA 3.b: To assess the duration of performance boost following a bout of meditation. This duration may change over time with experienced meditators. Several reports have suggested that there is substantial re-wiring and even increased cortical thickness in some regions of the brains of long-term meditators. Therefore, we will test whether the performance boost is different or lasts longer in the case of practicing experienced meditators as compared to novice or non-meditators. We also plan to observe and analyze the differences in brain activity patterns in experienced and novice meditators using raw EEG data. SA 4: Open source analysis tools (Kumar): Rationale: Typically software developed in the lab for the lab is not easily used elsewhere. Dr. Kumar at KSU is an expert in human-computer interactions and computing in education. He and his HBCU students will focus on implementation of the algorithms developed in Specific Aims 1-3 into user friendly interactive software tools that can be used for further investigation of these phenomena at other laboratories that may not have the necessary signal analytic and computational resources, and for educational purposes with the aim of increasing awareness and interest in STEM based education.
Effective start/end date8/1/158/31/21


  • University of Rhode Island: $984,056.00


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.