Center for Neural Engineering and Neuroscience Seminars Spring 2007

Wednesdays, 4pm, 108 Wartik
also simulcast to Hershey, PA campus, Room CG623
Also available as WebCast, or archive (contact Jean Brooks jeb4@engr.psu.edu ).

Spike-to-spike correlation and neural coding: basic roles for membrane and circuit dynamics

Abstract:

Correlations among neural spike times are ubiquitous, and questions of how these correlations develop, and of the impact they have on the neural code, are central in neuroscience. Their analysis also poses rich theoretical and mathematical problems. We address two of the most basic ones here.
First, we ask: How do correlations among different neurons depend on the cells' operating range – their rate and regularity of spiking? We use both linear response calculations and in vitro experiments to show that correlations between pairs of neurons vary sharply with their firing rates, almost universally. We illustrate the consequences via Fisher information, which quantifies the accuracy of encoding.
Next, we ask: How do correlations among different trials depend on architecture of neural circuits? (Here, the same stimulus is received by the circuit on each 'trial.') We take a first step toward the answer by identifying a surprising role for some, but not all, feedback connections in creating unreliable (and hence decorrelated) responses, a phenomenon which we quantify via Lyapunov exponents.

Note: This seminar will be available as an open broadcast to the Penn State Community at: http://live.libraries.psu.edu/mediasite/viewer/

Dr. Brown will present a Journal Club on his recent work, Optimal Deep Brain Stimulation of the Subthalamic Nucleus - a Computational Study, at 0900 on January 24 in 216 EES Bldg - please email jeb4@engr.psu.edu if you would like to view the manuscript and attend.

Songbirds and Synfire Chains

Abstract:

Many species of songbirds do not sing instinctively but learn their songs by a process of auditory-guided vocal learning that starts with a kind of babbling that converges over several months and through tens of thousands of iterations to a highly precise adult song. How the neural circuitry of the songbird brain recognizes, generates, and learns temporal sequences related to song are important questions for neurobiology and also interest an increasing number of physicists with backgrounds in nonlinear dynamics, acoustics, biophysics, and condensed matter physics. I will discuss some of the fascinating questions posed by recent experiments on songbirds. I will then discuss a theoretical model known as a synfire chain that my group [1] and others have invoked as a possible explanation for some features of the experimental data. [1] “Stable Propagation of a Burst Through a One-Dimensional Homogeneous Excitatory Chain Model of Songbird Nucleus HVC”, Meng-Ru Li and Henry Greenside, Physical Review E 74:011918 (2006).

Spike timing amplifies the effects of electric fields on brain function: a potential role for field oscillations

Abstract:

Despite compelling phenomenological evidence that small electric fields can affect brain function, a comprehensive and experimentally verified theory is currently lacking. Here we demonstrate a novel mechanism by which the non-linear properties of single neurons 'amplify' the effect of small electric fields: when concurrent to supra-threshold synaptic input, small electric fields can have significant effects on spike timing. The effects of fields on spike timing are amplified with decreasing synaptic input slope and increased cell susceptibility (mV membrane polarization per field amplitude). We confirmed these predictions experimentally using CA1 hippocampal neurons in vitro exposed to static (DC) and oscillating (AC) uniform electric fields. Our results provide a precise mechanism for a functional role of endogenous field oscillations (e.g. gamma) in brain function, and introduce a novel framework for considering the effects of environmental fields and design of low-intensity therapeutic neuro-stimulation technologies.

Translational Neuroengineering and Devices for Epilepsy

Abstract:

Of the world's 60 million people with epilepsy, approximately 1/3 have seizures that cannot be controlled by medications. Resective epilepsy surgery was conceived to cure these medically refractory patients, though there is now a growing realization of the limitations of this therapy outside of patients with clear focal lesions or mesial temporal sclerosis. More than 25% of individuals with epilepsy cannot be controlled by any available therapy, including surgery. Antiepileptic devices hold significant hope for these individuals, and for those whose seizures are controlled at the cost of significant side effects. Like epilepsy surgery, optimal benefit from devices will depend upon localizing epileptic networks, understanding seizure generation, its underlying mechanisms, and developing techniques to pre-empt or arrest seizures. New research applying engineering techniques to analyzing clinical and research signals generated by epileptic brain hold tremendous promise to tackle these fundamental issues. They are already giving us more insight into how seizure generation begins, and making us question clinical methods and approaches that have been used for the past 35 years. This talk will assess the major challenges to treating medically refractory epilepsy, explore recent work in medical devices applied to treating seizures, and discuss new work in Translational Neuroengineering applied to epilepsy.

NOTE: This seminar is being broadcast to HERSHEY room CG623 for colleagues at the College of Medicine,

and will be available as an open broadcast to the Penn State Community at: http://live.libraries.psu.edu/mediasite/viewer/

A possible mechanism for the efficacy of subthalamic nucleus deep brain stimulation

Abstract:

The delivery of high-frequency stimulation (deep brain stimulation, or DBS) to the subthalamic nucleus (STN) or other target areas has become a widely used therapeutic option for the treatment of Parkinson's disease (PD) and other neurological disorders. The mechanisms underlying its effectiveness, however, remain unclear and under debate. Because lesions of the STN and DBS of the STN lead to similar therapeutic outcomes, it has been hypothesized that DBS blocks STN activity, causing an effective lesion. More recent studies, however, have suggested that DBS activates areas downstream from its target site. In the case of PD, this leads to the paradox that although increased inhibitory basal ganglia output is associated with motor symptoms, STN-DBS may increase inhibitory output further yet relieve these symptoms.

In this talk, I will present work done in a computational model that explains a possible resolution of this paradox. A key aspect of this explanation is that STN-DBS may regularize basal ganglia output, eliminating the pathological rhythmicity associated with PD. The work that I will discuss provides the first detailed demonstration of how such regularization could normalize motor processing, through a restoration of thalamocortical relay in areas targeted by the basal ganglia. In addition to a purely computational exploration of the mechanism involved, I will present results that arise when basal ganglia outputs recorded from MPTP monkeys are incorporated into a computational model. These results show significant distinctions between thalamocortical relay under parkinsonian conditions without DBS, with sub-therapeutic DBS, and with therapeutic DBS.

Note: This seminar will be available as an open broadcast to the Penn State Community at: http://live.libraries.psu.edu/mediasite/viewer/

Oscillatory synchronization and information coding in the olfactory system

Abstract:

The olfactory system maps complex and high dimensional olfactory stimuli (odors) into unique and reproducible ensembles of neuronal activity. This mapping includes multilevel processing and involves complex strategies for the efficient encoding of information. In the olfactory system of insects, dense and dynamic odor representations generated in the antennal lobe (functional analog of the olfactory bulb) are transformed into sparse and specific patterns of neuronal activity in the mushroom body (analog of olfactory cortex). Remarkably, odor representations in the mushroom body remain sparse over thousand-fold changes in odor concentration, a feature potentially useful for storing and retrieving memories. Drawing on results obtained with biophysical network models of the olfactory system, I will discuss intrinsic and circuit properties that contribute to encoding olfactory information at different levels of odor processing, and the role of the intrinsic dynamics of the olfactory system in optimizing odor representations. I will also present a novel hypothesis that may reveal how competition between excitation and inhibition in the olfactory system creates a gain control mechanism for maintaining the stability and sparseness of neural codes for odors across broad ranges of concentrations.

A distributed spike-timing code in the primary somatosensory cortex and its enhancement by neuromodulators.

Abstract:

Single-trial response classification methods using information theoretic measures have been used to understand how neurons encode sensory information. We developed a peri-stimulus time histogram (PSTH)-based classification method to investigate the role of temporal precision in the encoding of sensory information by large populations of single neurons and to clarify the role of receptive field excitatory surrounds for cortical processing of somatosensory information. Our results demonstrate that the timing of the spikes plays a significant role in encoding stimulus location. In addition, although most of the information about stimulus location is preserved in the first spike after the stimulus, when a second spike occurs, it carries more information per spike than the information in the first spike. Moreover, not only trials correctly classified but also the trials incorrectly classified convey information about stimulus location with a similar temporal precision, suggesting that trials are incorrectly classified only due to the small sample of neurons. Importantly, when cells were divided into subpopulations with the same Principal (or Center) Receptive Field (PRF), the subpopulation could discriminate locations presented to the surround. Furthermore, the precise timing of spikes was relevant only for stimulation of the surround and not for stimulation of the center. These results suggest that: (1) the spatiotemporal structure of the excitatory receptive fields of S1 infragranular neurons allows cortical columns/segregates to ‘know’ when other columns/segregates are being activated; (2) surround responses are solely responsible for the transformation of spatial information (stimulus location) into spike-timing patterns in the forelimb somatosensory cortex of the rat. Finally we show that neuromodulators, such as serotonin, can increase the amount of information represented by the population by improving the acuity of the response and, therefore, the rate of information processing in the primary somatosensory cortex is state dependent.

Short bio: Karen Moxon, Ph.D. is an Associate Professor in the School of Biomedical Engineering at Drexel University and holds a joint appointment in the College of Engineering, Department of Electrical and Computer Engineering and College of Medicine, Department of Neurobiology and Anatomy. She received her BS in Chemical Engineering from University of Michigan, Ann Arbor and her Master’s and PhD Degrees in Aerospace Engineering from the University of Colorado, Boulder. Her lab studies how populations of neurons encodes sensory information and coverts it into appropriate motor output using computational modeling, recording of large numbers of single neurons and by using neurorobotic control paradigms.

Note: This seminar will be available as an open broadcast to the Penn State Community at: http://live.libraries.psu.edu/mediasite/viewer/

Brain-computer interfaces (BCIs) for communication and control

Abstract:

Brain-computer interface (BCI) research seeks to develop new augmentative communication and control technology for human patients with severe neuromuscular disorders, such as amyotrophic lateral sclerosis (ALS), brainstem stroke, and spinal cord injury. The goal is to give these users, who may be totally paralyzed (“locked in”), basic communication and control capabilities so that they can express their desires to caregivers or even operate word processing programs or neuroprostheses. Current BCIs determine the intent of the user from scalp-recorded electrical brain signals (EEG), or from electrodes surgically implanted on the cortical surface (ECoG) or within the brain (Neuronal action potentials or local field potentials). These signals are translated in real time into commands that operate a computer display or other device. Successful operation requires that the user encode commands in these signals and that the BCI derive the commands from the signals. Thus, the user and the BCI system need to adapt to each other initially and continually to ensure stable performance. This dependence on the mutual adaptation of user to system and system to user is a fundamental principle of BCI operation.

BCI research at the Wadsworth Center focuses on non-invasive EEG-based BCI methods and on moderately invasive ECoG-based methods. We have shown that patients with motor disabilities can learn to control amplitudes of EEG sensorimotor rhythms and can use this control to move a cursor rapidly and accurately in one or two dimensions. Current EEG-based multidimensional control is comparable in speed and accuracy to that reported using implanted electrodes (e.g., compare the videos at the first two web sites listed below). We are now going on to develop sequential “reach and grasp” movement control. Parallel studies are underway using ECoG signals recorded from people implanted temporarily with electrode arrays on the cortial surface prior to epilepsy surgery. Initial studies suggest that ECoG should be able to provide communication and control that is substantially faster and more precise than theat currently possible with EEG

At the same time, we are engaged in an effort to demonstrate that a simplified EEG-based BACI system can function reliably in the homes of patients with severe disabilities, can provide them with communication functions that are useful to them in their daily lives, and can do so without requiring excessive ongoing technical support. This simplified BCI system uses our standard general-purpose BCI software platform, BIC2000 (which we have provided to more than 100 other labs throughout the world, for research purposes only). Our BCI home system has a simplified electrode cap and can use either sensorimotor rhythms or P300 evoked potentials as the EEG signal features that provide control. It uses a highly flexible, sequential menu-based format that can be configured for the needs and capacities of each user (e.g., for word-processing, environmental control, entertainment access, e-mail, etc.).

We have begun to provide this BCI home system to a selected group of severely disabled users whose current communication methods are extremely limited and/or unreliable. We seek to determine: to what extent they use the BCI system in their daily lives; to what extent we can minimize the need for ongoing technical support; and to what extent the BCI system improves quality of life for these users and their families and caregivers. The first user is a highly productive scientist with ALS who has only eye-movement remaining. He has found the Wadsworth BCI system to be superior to his eye-gaze system. For the past year, he has been using the BCI up to 6-8 hours/day for writing e-mail and for other purposes. We have recently provided two other people similarly disabled by ALS with the Wadsworth BCI home system.

our goal over the next 2-3 years is to further improve the capability and reliability of the BCI home system, to provide it to an initial user group of 20-30 people with severe disabilities, and to show that it is useful to them and improves their quality of life. We intend to develop a network of clinical sites, each of which will manage the BCI use of its own patients with assistance from the Wadsworth BCI group as needed. Over the next 5 years, we hope to establish a self-sustaining non-profit entity that will make BCI technology widely available to those with severe disabilities and will continue to serve as a pipeline for the initial clinical validation and subsequent dissemination of new BCI technologies (e.g., ECoG) and applications.

The Fusion of Touch and Position Signals for the Localization of Sensation in Body-Centered Coordinates

Abstract:

The sensations of touch that occur when an animal encounters an object in front of its head as opposed to its side can be similar, but the two sensations inform the animal about very different things in its environment. The translation from passive contact with an object to knowledge of where the object is located requires that the nervous system keep track of the position of the animal's body and sensors as they move. We address the underlying mechanism for this process at the psychophysical through electrophysiological levels using the vibrissa system in awake, trained rats as our model. At the level of psychophysics, we found that rats with only a single vibrissa can combine touch and movement to distinguish the location of objects that vary in angle along the sweep of vibrissa motion. The patterns of this motion and of the corresponding behavioral responses show that rats can scan potential locations and decide which location contains a stimulus within 150 ms. This interval is consistent with just one to two whisks and provides constraints on the underlying perceptual computation. Our data argue against strategies that do not require the integration of sensory and motor modalities.

At the level of electrophysiology, we find that two forms of vibrissa-based signals are encoded as spike trains. One is a touch-based signal that yields a prompt spike upon contact of the vibrissa with a stimulus. The second is a motion-based signal, for which spikes are generated at a particular phase in the whisk cycle as animals whisk without contact. Interestingly, we find units whose probability of spiking depends on the conjunction of both touch and motion. These units spike only if contact occurs at a specific phase in the whisk cycle and thus provide the neurological basis for encoding touch in body-centered coordinates.

Position-dependent spike rates not only inform the rat of the location of an object in body-centered coordinates, but may trigger a change in their spatial orientation and/or whisking strategy. Consistent with this viewpoint is the observation that primary vibrissa motor cortex can rapidly control whisking in aroused animals. The computational underpinnings for this process are under investigation.

Motor Cortical Encoding of Planned Hand Orientation in a 3-D Reach-to-Grasp Task

Abstract:

In the pursuit of developing thought controlled neuroprosthetic systems it is critical to understand whether, where and how the neuronal populations in the cerebral cortex encode the control commands for hand movement and orientation in space. We designed an experiment to investigate the location and distribution of neurons in the motorcortex exhibiting a high level of correlation of spike activity patterns and hand movement in 3D space during a reach to grasp task. We recorded 2235 motor cortical cells, of which 1676 cells were found to be task-related but correlate with different movement control parameters. In a further experiment to investigate how robust these coding patterns are under different task conditions, we introduced a sudden change of target orientation 70 msec after movement onset. The task was performed under three perturbation conditions: fixed orientations, randomly changed orientations, and predictably changed orientation.

The data showed that in the motor cortex a portion of neurons could code hand orientation independent of movement direction. At the same time, cells encoding hand orientations and cells encoding movement directions coexist in the same region of the motor cortex, which indicates there probably exists a common pathway in M1 that controls both parameters. The perturbation results also show that feedback control is important to the execution of the movement, but CNS tends to optimize feed-forward planning when the disturbance information is predictable.