Randall K. Powers - US grants

The momenttomoment control of movement depends both upon the descending commands to the spinal cord and on activity in spinal reflex pathways. Although the reflex actions of muscle stretch have been extensively studied, much less is known about the reflex actions of muscle stretch have been extensively studied, much less in known about the reflex actions of muscle contraction. In particular, there has been very little work on the neuronal mechanisms and functional role of reflex effects directed from the contracting muscle to other muscles. The purpose of the proposed research is to provide a quantitative analysis of these contractionevoked reflexes and to investigate their afferent and interneuronal mechanisms. This general goal can be divided into 3 specific aims: 1. To describe the pattern of reflex effects produced by muscle contraction in the decerebrate cat (with or without an additional partial spinal section). Reflex effects produced by isometric contractions (via ventral root stimulation), stretch, vibration and electrical stimulation of muscle nerves will be compared with the responses of different types of muscle afferents to determine the afferent source of these effects. 2. To describe the response of Group 1 and force activated interneurons to muscle contraction, stretch, vibration and nerve stimulation. Quantitative comparisons of reflex effects with afferent and interneuron response will be used to identify those interneurons most likely to mediate contractionevoked reflexes. 3. To determine the specific output connections of Group 1 and force activated interneurons by spiketriggered averaging from neuron discharge into electomyographic activity in various extensor muscles.

"Nonlinear Systems Analysis of Spike Encoding in Montoneurons"

Individual nerve cells (neurons) transform the chemical and electrical signals they receive from other neurons into a series of electrical impulses or spikes. The frequency and pattern of these spikes forms the "neural code" by which information is transmitted throughout a network or system of neurons. Thus far, descriptions of the basic input-output transform of neurons, called spike encoding, have been largely limited to steady-state conditions in which the input signals that the cell receives are held constant. The objective of this proposed research program is to derive a general, quantitative description of spike encoding that will apply to both steady-state and dynamic conditions in which the input signals vary as they do under normal physiological conditions.

In the proposed experiments, the responses of mammalian neurons to brief injected current transients will be measured. The injected current transients are constructed to mimic real inputs as they appear in the cell body of a neuron. Non-linear systems identification procedures originally developed for electrical engineering applications will be used to characterize how these input signals affect the generation of spikes by the neuron. The advantage of this approach is that it yields a basic input-output function that is not as computationally complex as those derived from more detailed neuron models, but still accurately reproduces a wide range of neural behaviors. This compact input-output function can be then incorporated into the neural elements used in models operating at the network and systems levels, increasing the degree to which such models can accurately represent their biological counterparts.

DESCRIPTION (provided by applicant): All movements are executed as a result of graded activation of different muscles. Muscle activity is controlled by the activation of motoneurons in the brainstem and spinal cord. Each motoneuron drives the muscle fibers it innervates in a one-to-one fashion, thus forming a motor unit. Because muscle fiber action potentials are relatively easy to measure, motoneurons are the only CNS cells whose firing patterns can be readily quantified in human subjects. The cellular mechanisms that drive these firing patterns, however, can only be measured via intracellular studies in animal preparations. The goal of this proposal is to develop a sophisticated computer simulation platform to quantitatively link cellular data from animal preparations to firing pattern data in human subjects. Highly realistic models of human motoneurons will be implemented on field programmable gate arrays (FPGAs). We will employ these models to quantify our present state of knowledge about cellular mechanisms of human motoneuron firing patterns. The simulations will then be used to generate predictions for further experiments both in humans and animals, with the goal of identifying mechanisms underlying the severe deficits in firing patterns that occur in hemiparetic stroke patients. The overall hypothesis of this proposal is that these deficits in firing patterns are primarily due not to alterations in the synaptic input to motoneurons but instead to changes in their intrinsic electrical properties. Normally, motoneuron intrinsic properties are controlled by descending neuromodulatory inputs from the brainstem that release the monoamines serotonin (5HT) and norepinephrine (NE). Thus, changes in intrinsic properties may arise from changes in the input from the brainstem to the spinal cord. The proposal has three specific aims: 1) To develop highly realistic models of human motoneurons using a high-speed (FPGA) simulation platform in conjunction with automatic parameter search algorithms;2) To use these models to identify potential cellular mechanisms underlying changes in motoneuron firing patterns in hemiparetic stroke;and 3) To carry out new experiments in humans and animal models to test predictions developed in the Aim 2 model analyses. The results of these studies have the potential for substantial clinical impact. Drugs that mimic the effects of two important motoneuron neuromodulators, the monoamines 5HT and NE, have especially strong actions on these cells'properties. Thus, the proposed work will not only provide a new level of understanding of cellular properties of human motoneurons, but also guide development of therapeutic strategies to restore normal motoneuron discharge patterns in stroke patients. PUBLIC HEALTH RELEVANCE: Cerebral strokes commonly result in a number of movement deficits on the side of the body opposite the stroke (hemiparesis). The proposed research combines computer simulations with experimental recordings in hemiparetic stroke subjects and in animal models to determine the mechanisms underlying movement deficits following stroke. The proposed work will not only provide a new level of understanding of the cellular properties of the cells that drive muscle activity, but also guide development of therapeutic strategies to restore normal muscle activation in stroke patients.

DESCRIPTION (provided by applicant): All movements are controlled by the graded activation of different muscles. Muscle activation is controlled by the activation of motoneurons in the brainstem and spinal cord. Each motoneuron drives the muscle fibers it innervates in a one-to-one fashion. Muscle fiber action potentials are relatively easy to record in human subjects and it has long been appreciated that the firing patterns of motor units provide a unique window on the properties of motoneurons and their inputs. The goal of this proposal is to deduce the pattern of inputs to motoneurons based on an analysis of their output. The success of this reverse engineering approach depends upon two factors: (1) the existence of accurate models of the input-output properties of a population of motoneurons, and (2) the ability to extract the maximum amount of information from motoneuron output, based upon recording the activity of multiple, simultaneously-active motor units. We have recently developed a set of motoneuron models that can replicate several key features of the input-output properties of motoneurons with different intrinsic excitability. When these models are driven with different patterns of excitatory and inhibitory inputs they reproduce a number of the features of the discharge of human motor units during voluntary contractions. Previously, recording the behavior of multiple units required combining recordings taken across multiple sessions and subjects. New developments in recording and analysis techniques now make it possible to record from 10 or more motor units in each trial. This provides a rich data set that can be used to constrain the set of input patterns that give rise to a given pattern of output. The overall hypothesis of this proposal is that the firing patterns of a population of motoneurons are determined by the patterns of their synaptic inputs and the level of neuromodulatory drive, making it possible to deduce input patterns from the firing patterns of multiple motor units. We will test this hypothesi by comparing our estimates of input based on the reverse engineering approach with direct voltage-clamp measurements of the inputs in the same experimental preparation. There are three specific aims (1) to improve techniques for recording the simultaneous activity of multiple motor units, (2) to validate our reverse engineering approach by comparing synaptic input patterns that are predicted from recordings of motor unit discharge to those that are directly recorded from motoneurons, and (3) to determine the role of synaptic inhibition in shaping motor unit discharge patterns, and the ability of our reverse engineering approach to detect different patterns of inhibition. Once successfully developed in our animal preparation, the reverse engineering methods can be adapted for use in human subjects, allowing maximal use of the rich information available in motor unit firing patterns for understanding the structure of motor commands in humans in both normal and pathological states.