Mark Shelhamer - US grants

A robust and versatile capability for adaptive control of the vestibuloocular reflex (VOR) is essential for an organism to maintain optimal vision throughout life. Changes with development, aging, disease and trauma, demand mechanisms to detect and correct errors in performance. An understanding of such adaptive mechanisms bears on a fundamental problem in neuroscience -- motor learning -- and is also essential for accurate clinical diagnosis and the design of physical therapy programs. The long-term goal of this research is to understand how humans adapt to vestibular disorders, with the practical aim of developing better physical therapy programs for patients with vestibular disorders. The specific objective of this project is to learn more about the mechanisms underlying short-term -- minutes to hours -- VOR adaptation in normal humans. The emphasis is upon adaptive control of 1) otolith-ocular reflexes, 2) the phase of the canal and otolith-ocular reflexes, and 3) the torsional VOR. The error signals and contextual cues that lead to the expression of adapted responses will be evaluated. The relationship of adaptation of the VOR to the function of the ocular motor gaze-holding neural integrator, and to predictive pursuit mechanisms will also be investigated. Relatively little is known about these aspects of vestibular physiology, and each potentially bears on important issues related to vestibular adaptation, the error signals that drive it, and how adaptation can be promoted in patients. Eye movements will be measured using the magnetic field search coil technique. Otolith-ocular reflexes will be elicited in response to changes in the orientation of the head with respect to gravity (ocular counterroll), during eccentric rotation, which combines angular and linear acceleration, and in response to translation on a linear sled. Adaptation will be elicited using a visual-vestibular conflict paradigm in which the VOR is made to seem inappropriate by rotating or translating the visual scene (an optokinetic drum or a projected visual scene relative to the motion (rotation or translation) of the head). The results of such experiments will have important theoretical implications for basic vestibular physiology (and lend themselves to experimental neurophysiological study and mathematical modeling) as well as potential practical applications to clinical neurootology.

The investigators will develop a new device for the precise and convenient measurement of eye movements. The device is based on the existing and widely-used wired scleral search coil technique, in which an annular contact lens with a small coil of wire is placed on the eye, and its orientation determined from the magnitude of the current induced in the coil by external magnetic fields. The new device uses a similar scleral contact lens on the eye, but the new lens contains three resonant coils, and has no wires leading from these coils to the associated electronics. A small set of transmitter/receiver coils placed near the eye generates magnetic fields that induce current in the eye-mounted coils, and in turn detect re-radiated energy from the resonant eye coils. The orientation of the lens and coils is determined from the magnitudes of the received signals. Development of the new device involves the fabrication of three coils on a single lens, resonant at three different frequencies, and implementation of associated signal processing algorithms for measuring the signals radiated by the coils and computing coil (eye) orientation.

This new method retains the conventional wired coil's benefits of high resolution and accuracy, large range of movement, and complete three-dimensional measurement of orientation. The largest drawback of the conventional system is the wires that lead from the eye coils to the signal processing electronics. By eliminating these wires, several very significant advantages result. Subject comfort is greatly increased, as the external wiring leading from the eye often irritates the eyelid and causes annoying stimulation of the eyelashes. A greater range of experiments can be performed, since the placement of the head within the external magnetic fields becomes less restrictive; head-free experiments, such as those involving locomotion and self-generated movements become much more practical. Perhaps most importantly, breakage of the external wiring (the most common source of failure in current systems) is avoided. The "wireless" measurement system will be of tremendous benefit to both the research and the clinical communities. In particular, vestibular and oculomotor patients should tolerate the new coils much better than existing ones, allowing easier and more convenient recording of precise eye movements in a larger segment of this population, for whom the conventional coils are often intimidating and uncomfortable.

DESCRIPTION (provided by applicant): Many physical systems exhibit patterns of behavior typified by phase transitions: as a "control parameter" varies, the behavior changes between two (or more) distinct stable states or modes. Recently, examples of such phase transitions have been demonstrated in physiological systems. For example, when subjects tap the index fingers of each hand, there are two stable modes: one in which the fingers tap in phase and one in which they tap in counter-phase. These two modes are robust (stable), and the behavior changes from one to the other as tapping frequency is systematically varied. This behavior exhibits a number of features common to phase transitions: each mode is robust to perturbations, there is increased variability when the control parameter (frequency) is near the transition point, and there is hysteresis such that the transition point is different when approached from above versus below (decreasing or increasing frequency). We will investigate oculomotor behavior for similar types of phase transitions. Mathematical modeling in the oculomotor field is quite sophisticated, yet analysis of phase-transition behaviors has not been attempted. The proposed work is thus an attempt to model an aspect of oculomotor behavior that has not been addressed. The modeling proposed here will have value in improving existing models, understanding observed behaviors, and predicting new ones. The research plan consists of three main efforts. First is the analysis of periodic saccades for evidence of multiple stable states, in analogy to the finger tapping work referred to above. Preliminary data suggests that phase transitions may indeed occur in this system: as target jumping frequency increases, saccades change from lagging the target (positive latency) to anticipating the target (negative latency), and there is increased variability near the transition frequency. Work on this system will involve further verification of this effect, examination of hysteresis and the response to perturbations, and formulation of mathematical models (and analysis of existing models) to reproduce the phase transitions. The second effort is a study of the interaction of saccades and the VOR (vestibulo-ocular reflex), looking for modes of coordination between the resulting smooth and fast eye movements as a function of frequency. Finally, we will carry out a survey of other vestibular and oculomotor behaviors and models for those which might exhibit phase transition behaviors. This project is high risk because the oculomotor system may avoid such phase transitions but rather exhibit smoothly graded behavioral changes, making phase transition models inappropriate. The work represents a new direction for the PI because his past work has not dealt with the unique mathematical formulations required for this study.

DESCRIPTION (provided by applicant): The investigators will develop a new device for the precise and convenient measurement of eye movements. The device is based on the existing and widely used wired scleral search coil technique, in which an annular contact lens with a small coil of wire is placed on the eye, and its orientation determined from the magnitude of the current induced in the coil by external magnetic fields. The new device uses a similar scleral contact lens on the eye, but the new lens has no wires leading from the eye coil to the associated electronics. Signals are induced in the eye coil, and detected from the eye coil, by a transmitter/receiver near the eye. The orientation of the lens and coil is determined from the received signal. An existing prototype system is the basis for further development of the new system, which will proceed along the following lines: 1) assessment of engineering design options with the existing system; 2) fabrication of new contact lenses with the appropriate coils embedded; 3) extension of the system to measure torsional eye movements as well as horizontal and vertical; 4) packaging as a stand-alone system that can be worn by the subject; 5) assessment of performance by video analysis of eye and lens motion. This new method retains the conventional wired coil's benefits of high resolution and accuracy, large range of movement, and complete three-dimensional measurement of orientation. The largest drawback of the conventional system is the wires that lead from the eye coils to the signal processing electronics. By eliminating these wires, several very significant advantages result. Subject comfort is greatly increased, as the external wiring leading from the eye often irritates the eyelid and causes annoying stimulation of the eyelashes. A greater range of experiments can be performed, since the placement of the head within the external magnetic fields becomes less restrictive; head-free experiments, such as those involving locomotion and self-generated movements become much more practical. Perhaps most importantly, breakage of the external wiring (the most common source of failure in current systems) is avoided. The "wireless" measurement system will be of tremendous benefit to both the research and the clinical communities. In particular, vestibular and oculomotor patients should tolerate the new coils much better than existing ones, allowing easier and more convenient recording of precise eye movements in a larger segment of this population, for whom the conventional coils are often intimidating and uncomfortable.

DESCRIPTION (provided by applicant): This study addresses adaptation to the unique environment of parabolic flight. Experiments will take place in a NASA aircraft that flies parabolic trajectories which make transitions between 0 g and 1.8 g approximately every 25 sec. This unusual motion environment, with changes in g level not normally experienced, presents a remarkable challenge to the body's adaptive processes. Indeed, many first-time flyers experience disorientation and motion sickness. Nevertheless, adaptive processes prevail and on the second and subsequent flights the experience is much more pleasant, indicating that significant adaptation has occurred in a very short time and after a very brief exposure (approximately one hour of parabolic maneuvers per flight). We propose to study the adaptive processes that take place in this situation. A battery of tests of vestibular and oculomotor function will be performed during the two different gravity levels in flight. The tests span a range of neural processes from low-level reflexive through higher-level cognitive, and are aimed in particular at otolith information (since these vestibular organs sense linear accelerations, including gravity) and vision (since visual information can in many cases substitute for vestibular information about self-orientation). The tests are: 1) ocular counterrolling (OCR), a reflexive torsional movement of the eyes in relation to head tilt in a gravity field as sensed by the otolith organs; 2) translational VOR (TVOR), a compensatory eye movement made in response to head translations that also depends on processing of otolith signals; 3) pitch angular VOR, which involves convergence of otolith and canal signals; 4) vertical ocular alignment, which is affected by gravity level; 5) subjective visual vertical, based on multi-sensory integration of orientation cues; 6) roll vection, which addresses the interaction of visual and otolith cues. Responses from new and from experienced flyers will be compared, and the progression of responses over consecutive flights will be assessed as adaptation develops. The tests address two competing hypotheses of how adaptation to parabolic flight is achieved: 1) there is adaptation to each separate gravity level (context-specific); 2) adaptation is more generalized to the overall flight experience (implying a non-g-specific change in sensory weighting). The information to be gained from this study of adaptation may lead to better understanding of the range of possible adaptive mechanisms, and might help us to determine those mechanisms which act most rapidly and effectively (as they do in parabolic flight) as candidates for use in rehabilitation of vestibular patients.

DESCRIPTION (provided by applicant): This project addresses two types of eye movement (saccades, OKN), with emphasis on timing in response to visual stimuli. Saccades are rapid movements made when scanning a visual scene. Saccades to periodically-paced targets have a characteristic latency: as pacing rate increases saccades become more anticipatory. OKN (optokinetic nystagmus) occurs with wide-field motion of a scene, and is an indicator of vestibular and oculomotor function. During OKN, the eyes track the field, and intermittently reset their position by moving rapidly in the opposite direction. The eye position waveform looks like a fluctuating sawtooth. OKN is not periodic; models typically rely on statistical characterization of the variations, and recent work indicates that the dominance of random vs. deterministic dynamics makes this reasonable. If a subject actively follows the scene ("look nystagmus"), OKN is different from when a subject simply stares straight ahead ("stare nystagmus"). This may reflect different degrees of reflexive and volitional control. In the first sub-project, scaling properties of periodically-paced saccades will be studied. Many systems have a characteristic scaling behavior, such that variability increases as a power-law function of sequence length, reflecting a type of statistical long-term memory. Scaling of saccade latencies will be assessed with a parametric procedure, and verified graphically. The goal is to assess scaling over the range where saccades are reactive or predictive (latencies positive or negative, respectively). Evidence is that scaling behavior in the predictive regime is different from that in the non-predictive ("reactive") regime. This relationship can link reflexive/volitional behaviors to predictive/reactive behaviors. Evidence is also presented for altered scaling in cerebellar patients. The second sub-project will apply similar analyses to OKN fast-phase intervals. Evidence indicates that scaling is altered depending on the degree of volitional vs. reflexive behavior. This volitional/reflexive mix can be altered through stimulation of different types of OKN (horizontal vs. torsional, "look" vs. "stare"), and used to verify the relationship of scaling to behavior. The results will be widely applicable to physiologic systems that can be described by statistical scaling laws, and to systems which exhibit combinations of reflexive and volitional behavior and prediction. Since scaling changes with pathology, new means to assess vestibular or oculomotor pathology may result.

Human behavior is fundamentally predictive. One way to see this fact is by considering our ability to make fast and accurate movements that anticipate upcoming changes or events in the environment. A tractable domain in which to study predictive behavior is eye movements. Saccades are rapid eye movements that everyone uses throughout their waking lives to scan their visual environments in order to maintain and update perceptual representations. Saccades are known to be ballistic, which means that once a saccade is initiated, its course cannot be altered. Therefore saccades must typically predict the near future location of visual stimuli so that the eyes land on their intended targets. In previous studies it was found that the nervous system holds information about the sequential history of eye movements in order to time future predictive eye movements. Intriguingly, the statistical nature of this memory is long-range correlated, and long-range correlations are mysteriously ubiquitous to many biological and complex systems found throughout nature.

With support of the National Science Foundation, Dr. Shelhamer will study nature of these long-range correlations in eye movements. In particular, he will examine the pattern of endpoint locations of saccades, and the characteristics of their spatial correlations. The experiments are designed to determine how far the correlations extend in time and space, how they differ between predictive and reactive saccades, and how they change with learning. An important aspect of the latter issue is that stronger correlations may be related to better adaptive capabilities. Motor adaptation is common to many neuromuscular systems, and the lessons learned in this study may help to increase the effectiveness of motor training and rehabilitation programs.

DESCRIPTION (provided by applicant): Humans have a remarkable ability to make predictive movements in order to overcome delays in neural processing, and to adapt to changes due to aging and disease. In each case, proper functioning relies on the detection and processing of sensory error information in order to properly program future behavior. We propose a new approach to presenting error feedback, to improve prediction and adaptation. When presented with periodically paced visual targets paced at about 1 Hz, normal subjects naturally make predictive saccades - they are triggered before visual feedback from a given target, with latencies of -100 to +100 msec. Sequences of predictive saccades are correlated: performance of past saccades is stored and taken into account in the timing of subsequent saccades. Initial evidence in patients with cerebellar deficits is that they have deficiencies in making these predictive movements, and that this is due at least in part to deficiencies in monitoring, storing, and processing the errors of previous movements. The ultimate goal of this research program is to present saccade error information to these patients in new ways, to help them improve predictive-saccade timing and accuracy. Our approach is to augment the normal visual error from each saccade with auditory information or with visual information in a different form, and further to present error information that has been accumulated over several previous trials rather than just the most recent. The research plan has four aims. First, determine the effectiveness of augmented feedback for control of predictive-saccade timing. We will supply augmented auditory feedback of timing error to the subject on each trial, to drive timing error to a desired value and to decrease its variability. This feedback is in the form of a beep generated when a saccade is generated or when its timing falls within a desired range. Second, increase correlations between consecutive predictive saccades. These correlations occur naturally in normal subjects, and reflect the fact that previous performance is used to program subsequent movements. We will determine if there is a performance advantage to these correlations, and attempt to increase their extent by providing error feedback based on timing error accumulated over several previous saccades rather than the single preceding one. Third, use similar methods to control endpoints (amplitudes) of predictive saccades, using error feedback of position errors. Fourth, use similar methods to improve the ability to adapt to a double-step stimulus: a visual target moves as the eyes approach it, and after a period of adaptation the eyes make a saccade to the displaced target position when presented with a target at the initial location. By providing augmented feedback based on displaced target position, we hope to improve the rate and extent of this adaptation. All procedures will be performed on normals and cerebellar patients with deficits in prediction and adaptation. The methodology has its ultimate usefulness in motor learning, thus the final aim has the goal of improving adaptive capabilities of cerebellar patients, which can have profound impact on rehabilitation programs. PUBLIC HEALTH RELEVANCE: Impaired ability to make predictive movements, and to adaptive to changing conditions, can seriously impact health - especially if these largely automatic processes then require conscious processing. Impaired prediction adversely affects the ability to generate appropriate motor actions in anticipation of upcoming requirements, and impaired adaptation affects that ability to adjust for changes due to aging and disease. Procedures that train the brain to compensate for deficiencies in these areas can be of great benefit in programs of neural rehabilitation.

In this project, the investigators will develop a set of computational tools to investigate long sequences of continuous data from physiological systems. One interesting property exhibited by these data sets is called "fractal scaling." This means that the data are "self-similar" on different time scales, so that short sections of data, when rescaled properly, resemble larger sections of data and this scaling holds true over a large range of time intervals. This is a defining feature of a fractal. Another fractal property is that the data exhibit fluctuations on many different time scales. Not all physiological data sets show this attribute. Some show instead a particular time scale at which the variability is most prominent. Many analytical and computational tools to study fractals have been developed over the last two decades. However, the large number of these tools makes it difficult for many investigators to make the proper choice when analyzing any given data set. Furthermore, different computational tools are susceptible to different types of artifacts and different tools are valid for different data types. Erroneous or misleading results can be obtained when the tools are improperly applied. It is very difficult for a non-expert in fractal mathematics to understand and keep track of all of these issues, which can inhibit progress in the investigation of many physiological and behavioral systems. In this project, the research team will assemble a set of established computational tools for fractal time-series analysis, apply them to a variety of data sets with known and unknown properties (i.e., artificially generated data and real physiological data), and determine when different tools break down, when they are most effective, how to detect when erroneous results are being generated, and how to interpret the results. This will provide increased confidence in the application of these tools and encourage their widespread use.

Fractal fluctuations are common in the behavior of biological systems. They have been found in heartbeat intervals, stride intervals, firing rates of neurons, human reaction times, and many other cases. However, the source and significance of these characteristics are unclear. They could arise from fractal properties of ion channels in neurons and have no special role or relevance. On the other hand, they could be deliberately produced at the behavioral level and confer a physiological benefit. By developing and making freely available a standard set of computational tools that implement the most modern algorithms for assessing data for fractal properties, the current project can further our understanding of how sensorimotor systems organize, how behaviors are produced, how sensory information is processed and how perceptions emerge.