Marc A. Maier

Precision in isometric precision grip force is reduced in middle-aged adults.

We investigated age related changes in the control of precision grip in 29 healthy adults spanning early adulthood to middle age (21–67xa0years). Subjects performed a visually guided, isometric precision grip ramp-and-hold force-tracking task. Target force levels were 3, 6, and 9xa0N. Precision and performance of force regulation was quantified. Larger errors were made during the ramp than during the hold phase. Age correlated positively with the amount of error at the lowest (3xa0N) force level in both phases. Force onsets were systematically earlier in middle-aged subjects and the average slope of the force during the ramp decreased with increasing age. The results show that precision during low grip force control decreases already during middle age and those subjects may modify their force generation strategies to compensate for early and subtle degenerative changes in the motor system before decline in grip strength is apparent.

Modeling of the residual capability for people with severe motor disabilities: analysis of hand posture

People with severe motor disabilities use mainly their residual motor capability for the use of technical aids, and for the control of input devices to technical aids. This paper describes our work on characterizing the motor capability of the upper arm for patients with severe motor disabilities. This work is a continuation of a project aimed at modeling the arm posture of quadriplegic patients using STS (Spatial Tracking System) and at analyzing the compensatory strategies developed by hemiplegic patients while accessing physical interfaces for technical aids [5]. Here we report work undertaken for analyzing the posture of the hand: we have developed two calibration methods for the Cyberglove and compare their utility and ergonomics in applications on patients with motor disabilities. The first type of calibration proceeds sequentially and takes into account one joint after the other (of the hand and each digit), whereas the second procedure is based on a few key postures calibrating several joints at once. To compare the precision of both methods, four healthy subjects participated in experiments using the Cyberglove. We show that the first type of calibration is more accurate but takes longer, whereas the second is less accurate but shorter. This trade-off might be acceptable for assessing the manual workspace in patients with motor disabilities. In particular, excessive muscular fatigue and limited dexterity are decisive factors for choosing the calibration by key postures in patients. We applied the calibration by key postures to three myopathic patients and individually quantified their restricted manual working space.

A generic neural network for multi-modal sensorimotor learning

Abstract A generic neural network module has been developed, which learns to combine multi-modal sensory information to produce adequate motor commands. The module learns in the first step to combine multi-modal sensory information, based on which it subsequently learns to control a kinematic arm. The module can learn to combine two sensory inputs whatever their modality. We report the architecture and learning strategy of the module and characterize its performance by simulations in two situations of reaching by a linear arm with multiple degrees of freedom: (1) mapping of tactile and arm-related proprioceptive information, and (2) mapping of gaze and arm-related proprioceptive information.

Temporal Processing in primate motor control: relation between cortical and EMG activity

We investigated spatio-temporal information processing in the primate motor system. Corticomotoneuronal (CM) cells provide monosynaptic excitatory connections from motor cortex to spinal motoneurons and contribute causally to the time-varying electromyogram (EMG) of their target muscle. A multilayer perceptron (MLP) was used to evaluate the transfer function between neural activity of single CM cells and their target muscle EMG, using data from in-vivo recordings in primate motor cortex. For an optimal MLP performance, i.e., minimal error between recorded target EMG and MLP-derived EMG, the CM cell input period had to span the latency observed between CM cell peak activity and EMG peak activity. We argue that the same spike train may code two types of information: 1) rate coding within the input window accounted for large-amplitude variations in the EMG signal and 2) temporal coding within a window of 40 ms just prior to the EMG output signal accounted for EMG variations of small amplitude. The transfer function of the MLP, thus, combines rate and temporal coding and suggests that CM cell output may also combine these two forms of coding. We predict that mutual constraints of rate and temporal coding would, however, would limit the CM output to code for particular temporal profiles of EMG, possibly adapted to bio-mechanical constraints.

Kinetic DTI of the cervical spine: diffusivity changes in healthy subjects

IntroductionThe study aims to assess the influence of neck extension on water diffusivity within the cervical spinal cord.MethodsIRB approved the study in 22 healthy volunteers. All subjects underwent anatomical MR and diffusion tensor imaging (DTI) at 1.5xa0T. The cervical cord was imaged in neutral (standard) position and extension. Segmental vertebral rotations were analyzed on sagittal T2-weighted images using the SpineView® software. Spinal cord diffusivity was measured in cross-sectional regions of interests at multiple levels (C1–C5).ResultsAs a result of non-adapted coil geometry for spinal extension, 10 subjects had to be excluded. Image quality of the remaining 12 subjects was good without any deteriorating artifacts. Quantitative measurements of vertebral rotation angles and diffusion parameters showed good intra-rater reliability (ICCu2009=u20090.84–0.99). DTI during neck extension revealed significantly decreased fractional anisotropy (FA) and increased radial diffusivity (RD) at the C3 level and increased apparent diffusion coefficients (ADC) at the C3 and C4 levels (pu2009<u20090.01 Bonferroni corrected). The C3/C4 level corresponded to the maximal absolute change in segmental vertebral rotation between the two positions. The increase in RD correlated positively with the degree of global extension, i.e., the summed vertebral rotation angle between C1 and C5 (Ru2009=u20090.77, pu2009=u20090.006).ConclusionOur preliminary results suggest that DTI can quantify changes in water diffusivity during cervical spine extension. The maximal differences in segmental vertebral rotation corresponded to the levels with significant changes in diffusivity (C3/C4). Consequently, kinetic DTI measurements may open new perspectives in the assessment of neural tissue under biomechanical constraints.

Recurrent neural networks of integrate-and-fire cells simulating short-term memory and wrist movement tasks derived from continuous dynamic networks

Dynamic recurrent neural networks composed of units with continuous activation functions provide a powerful tool for simulating a wide range of behaviors, since the requisite interconnections can be readily derived by gradient descent methods. However, it is not clear whether more realistic integrate-and-fire cells with comparable connection weights would perform the same functions. We therefore investigated methods to convert dynamic recurrent neural networks of continuous units into networks with integrate-and-fire cells. The transforms were tested on two recurrent networks derived by backpropagation. The first simulates a short-term memory task with units that mimic neural activity observed in cortex of monkeys performing instructed delay tasks. The network utilizes recurrent connections to generate sustained activity that codes the remembered value of a transient cue. The second network simulates patterns of neural activity observed in monkeys performing a step-tracking task with flexion/extension wrist movements. This more complicated network provides a working model of the interactions between multiple spinal and supraspinal centers controlling motoneurons. Our conversion algorithm replaced each continuous unit with multiple integrate-and-fire cells that interact through delayed synaptic potentials. Successful transformation depends on obtaining an appropriate fit between the activation function of the continuous units and the input-output relation of the spiking cells. This fit can be achieved by adapting the parameters of the synaptic potentials to replicate the input-output behavior of a standard sigmoidal activation function (shown for the short-term memory network). Alternatively, a customized activation function can be derived from the input-output relation of the spiking cells for a chosen set of parameters (demonstrated for the wrist flexion/extension network). In both cases the resulting networks of spiking cells exhibited activity that replicated the activity of corresponding continuous units. This confirms that the network solutions obtained through backpropagation apply to spiking networks and provides a useful method for deriving recurrent spiking networks performing a wide range of functions.

Cortical Reorganization Allows for Motor Recovery after Crossed Cerebrocerebellar Atrophy

The authors report the case of a 33‐year‐old woman who exhibited, at the age of 17, a left‐sided hemiplegia, which was followed by good motor recovery, though with a permanent deficit in fine finger movements. She had a widespread loss of neural tissue in the right hemisphere (crossed cerebrocerebellar atrophy), including (1) marked atrophy and thinning of the precentral and postcentral gyri; (2)widespread deep white matter destruction, including the corticospinal tract; and (3) crossed cerebellar atrophy. Except over the supplementary motor area (SMA), transcranial magnetic stimulation did not elicit motor evoked potentials in the affected hand. Nevertheless, during opening and closing of the affected hand, functional magnetic resonance imaging showed an activation of the lesioned primary sensorimotor cortex (SMC), as well as of the intact SMA and the parietal areas, but not of the ipsilateral motor areas. The authors speculate that recovery was achieved by a motor command generated in the SMC and the parietal cortex, passing through corticospinal axons originating in the SMA.

TempUnit: A bio-inspired neural network model for signal processing

We have developed and tested a novel artificial neural network for the processing of temporal signals. The working of the units (TempUnit) is based on the mechanism of temporal summation as observed in biological neurons. Contrary to traditional neural networks, the TempUnit optimizes its basis function by supervised learning. The model was tested on cortical and associated muscular (EMG) recordings from the behaving primate. The TempUnit showed a 2.3 times better performance in mapping spiking to EMG activity than a time delay multi-layer perceptron. The TempUnit model demonstrated correct capacities for inverse computation. Indeed, we calculated biologically compatible activities for 3 cortical neurons from EMG recordings. Data compression capacity of the TempUnit was tested on audio data and compared to the MP3 compression standard. For a similar reproduction quality, we found a compression rate 5 times higher than in MP3.