Zeynep Erim - US grants

9612191 Erim The objective of this study is to investigate the usefulness of two state-of-the-art methods in the analysis of electrical signals recorded from human muscles. These techniques are wavelet and time-frequency transforms, which have yielded favorable results in data compression, nonstationary signal analysis and noise removal; and fuzzy logic which has been successfully employed in decision- making and control fields. These methods will be applied to two main problems encountered in the analysis of signals recorded from muscles: a) the identification of the activities of individual muscle units represented in signals recorded via needle electrodes, and b) the analysis of the time-varying nature of signals recorded during dynamic contractions via surface electrodes attached to the skin over the muscle. The availability of powerful analysis tools will enhance the research in these fields geared toward gaining a better understanding of the control of muscles by the central nervous system, as well as the development of reliable diagnostic procedures to be used in the clinical environment. The initial year will be devoted to the implementation of the techniques proposed to be employed in the study, and to the application of these to specific signals recorded from muscles. It is hoped that this year will serve to provide the principal investigator with expertise in these new fields and help to establish her as an independent researcher. It is further hoped that the work carried out during this year will produce the preliminary results necessary to demonstrate the applicability of the mentioned techniques to the investigation of muscle signals, and that it will lead to the proposal of at least one full-blown study structured according to the findings and conclusions of the present study. ***

9973680
Erim
The purpose of the proposed work is to apply recent advances in Time/Frequency transforms to a specific problem in electrophysiology. The problem in question is that of observing the firing patterns of individual motor units (muscle fibers activated as a group) for the purposes of basic research, and for diagnosis and outcome assessment in rehabilitation in the clinical environment. The project aims to further and complete previous NSF-funded work that enabled the identification of the activities of motor units when their discharges did not overlap in time. The goal of this project is to develop a technique to decompose action potentials when they do overlap in time, using the cross-time-frequency transform.

In assessing muscle function by studying the firing activity of the individual motor units that make up the muscle, electromyographic (EMG) signals are typically detected via needle electrodes inserted into the muscle. Using the fact that each motor unit generates a distinct action potential waveform when it fires, it is possible to determine, from the needle EMG signal, the instances of firing for each motor unit within the pick-up area of the probe. This information is relevant in determining the strategies employed by the central nervous system in controlling the production of muscle force. This assessment is important both for basic research to understand the normal production of force, and for the clinical diagnosis of neuromuscular deficiencies. The accurate identification of motor unit firings is also important in ensuring a reliable representation of action potential shapes, which is useful in the diagnosis of dysfunctional muscles.

The decomposition of the needle EMG signal into the activities of single motor units becomes complicated when the action potentials of different motor units are similar, change shape with time or superimpose on each other. The problem becomes especially complicated when the action potentials that superimpose have an overlap in frequency content as well as overlapping in time. This project will apply cross-time-frequency distributions to the decomposition of superimposed waveforms. It is hoped that the availability of the representation of the time and frequency distribution of the signals will provide a better separation than is possible using just the time or frequency domain representation of the signal. The basic procedure will be based on pairing the superimposed waveform to be resolved with each of the identified action potential templates and computing the cross-time-frequency transform. The resultant cross-time-frequency transform will be compared to the auto-time-frequency representation of the action potential template in question to determine if the template is included in the superposition. The technique will be based on the notion that after proper alignment and filtering in the ambiguity domain, the cross-time-frequency transform will be similar to the time-frequency transform for the action potentials that were indeed included in the superposition, while considerably different than the time-frequency transform of the action potential when considering action potentials that were not involved in making up the superimposed waveform. Hence by comparing the cross-time frequency to the time-frequency representation, it will be possible to determine which action potentials were included in the superposition.

The ultimate goal is to develop a decomposition procedure that can correctly identify all motor unit firings, including those that are superimposed, from the raw needle EMG signal with minimal input from the operator, in order to greatly facilitate intramuscular EMG studies both for research and clinical purposes.

DESCRIPTION (provided by applicant): The candidate for this Mentored Quantitative Research Career Development Award was trained as a biomedical engineer and has performed experimental and modeling investigations to understand the principles underlying the regulation of muscle force, as well as technique development for the analysis, interpretation and modeling of the data. The mentoring team includes Dr. W. Zev Rymer, a leader in the fields of clinical neuroscience and rehabilitation engineering, Dr. Elliot Roth, an expert in stroke and spinal cord injury rehabilitation, and Dr. C.J. Heckman who is a foremost researcher of spinal cord circuitry and cellular mechanisms of spasticity. The candidate's long-term career goals are two-fold: to apply her experience with the normal operation of the neuromuscular system, to the understanding of the alterations and adaptations that take place in motor disorders so that better-informed interventions may be developed; and to employ the specific insults to the system associated with such disorders to further the basic understanding of the role and significance of the affected regions and pathways in the healthy operation of the system. The proposed research plan addresses the mechanisms underlying common firing patterns exhibited by human motor units. The first part will characterize the role of various central and peripheral sources in healthy subjects using techniques in which the candidate will receive training from distinguished experts of the field: EEG-EMG coherence analysis to be studied with Drs. Bernie Conway and Jay Rosenberg and TMS application with Dr. Leonardo Cohen. The second part of the project will investigate the joint behavior of motor units in two pathological conditions, stroke and complete spinal cord injury, drawing on the known pathologies to identify the role of various components in the generation of motor unit firing patterns. Dr. Christine Thomas will provide training on data collection procedures suited for spinal cord injured patients. The proposed career development program represents an invaluable contribution to the candidate's career providing her with new tools of investigation as well as a deeper understanding of the relevant neuroscience and rehabilitation issues. The Rehabilitation Institute of Chicago offers a uniquely suited environment for this application in terms of both the scientific and clinical expertise of its staff, and its patient population.

EEC-0649176
William Z. Rymer

The award provides support for a three-year REU Site at the Rehabilitation Institute Research Corporation. This proposal will establish a three-year Research Experiences for Undergraduates (REU) Site at the Sensory Motor Performance Program (SMPP) of the Rehabilitation Institute of Chicago. This program will enable 12 undergraduate students each year to spend 10 summer weeks participating in research projects focused around Neural Engineering. The program's primary goal is to provide hands-on research experience to students, and in, particular to introduce them to the field of Neural Engineering, with applications to human neurological disorders.

This REU site builds on a highly successful short-term research training program, which has been offered for the past 4 years. In particular, SMPP researchers, in collaboration with faculty from the Biomedical and Mechanical Engineering Departments of Northwestern University, have offered a successful and increasingly popular program entitled "Summer Internships in Neural Engineering" (SINE). Located within the research group of the nation's leading rehabilitation hospital, this REU program is in the position to offer unique training experiences that will motivate talented students to pursue graduate education and careers in Neural Engineering.

The program will afford women, underrepresented minorities, persons with disabilities and students from academic programs with limited research programs the chance to experience and contribute first-hand to state-of-the-art research activities in Neural Engineering and consider science and research as viable career options.