Context dependent changes of signal transfer at a central machine-brain interface. A study using mulitelectrode stimulation an d recording in barrel cortex of awake rats during active and passive touch.
Final Report Abstract
The field of cortical neuroprosthetics promises novel ways to aliviate symptoms of neurological diseases like sensory malfunction or paralysis of central origin. For sensory prostheses, current technology envisages to use intracortical microstimulation to activate small areas of sensory cortex to evoke artificial percepts. The problem with this method is the high interconnectivity of the cortex which, as we hypothesized, makes activity imprinted there susceptible to activity imprinted elsewhere or natural activity impinging from distant brain regions. If these modulatory influences are not known they must be expected to blur the precision of the imprint: indeed it is commonly known that neuronal responses to identical stimulation vary immensely from trial to trial. The present project, firstly, studied the interaction of multiple pulses on the percept of head fixed behaviorally trained rats. In line with our hypothesis, we found that a series of pulses invariably interact, such that the access of individual pulses to the percept is highly dependent on the pulse pattern applied. In a second approach, we experimentally manipulated the functional state of the cortex to be stimulated. We used the vibrissa representation of rat primary somatosensory cortex as a model system, in which whisking activity is known to set the functional state to higher spontaneous activity and lower sensory responses. Again supporting our hypothesis, we found that neuronal activity imprinted via microelectrodes is susceptible to modulation by the functional state, and we gained preliminary evidence that the detection threshold of microstimualtion pulses is lower in the active case. Thus the functional state of the cortex influences the subject’s perceptual reception of microstimulation. Spurred by these results we have begun a cooperative project (with Prof. M. Bogdan, Leipzig) that aimed to establish and test an algorithm that uses cutting edge machine learning technology to extract information about stimulation and/or functional context using simultaneous electrophysiological recordings from the cortex. This information is then used in real-time to adapt stimulation parameters such that the imprinted neuronal activity is stabilized (i.e. displays less trial to trial variability). Indeed, we found that local field recordings in the vicinity of the stimulation electrode contain significant amounts of information about the functional state of the cortex, such that the imprinted activity can be approximated to a given target activity with much higher precision as compared to ‘blind’ stimulation. At present these results were obtained in anesthetized animals. In a future cooperative project, which already received funding, the potential of this algorithm to deal with multielectrode recording and stimulation, and its ability to improve the percept of awake behaving rats will be explored in depth.
Publications
- (2007) Intracortical microstimulation in barrel cortex of awake, head-restraint rats: assessment of detection thresholds. Eur. J. Neurosci. 25:2161-2169
Butovas, S. and Schwarz, C.
- (2008) Direct and inverse solution for a stimulus adaptation problem using SVR. ESANN Proceedings 2008, pp. 397-402
Brugger, D., Butovas S., Bogdan, M., Schwarz, C. and Rosenstiel W.
