Neural Prosthetic Lab
The response to stimulation in axons has been well studied and theoretical models of the interaction are well supported by experimental data (the one-dimensional and uniform structure of axons facilitated the theoretical understanding). The response to stimulation in neurons of the CNS is considerably more complex. In these cases, stimulating electrodes are in close proximity to multiple regions of targeted cells (soma, dendrites, axons). Many of these regions have complex three-dimensional morphology and the underlying biophysical properties can also be non-uniform (e.g., the sodium channel density in different regions can vary considerably). We study how single neurons and populations of neurons respond to electric stimulation with the goal of understanding the factors that modulate activation. By understanding the fundamental principles of neuronal activation, we strive to develop more effective stimulation methods.
Retinal degenerative diseases such as age-related macular degeneration (ARMD) and retinitis pigmentosa (RP) are the leading cause of blindness in the United States and other industrialized countries. These diseases lead to a loss of photoreceptors, the neurons primarily responsible for transducing light into an electrochemical signal. Importantly however, retinal neurons downstream of photoreceptors do not degenerate, raising the possibility that artificial stimulation can be used to restore some form of vision.
The viability of this approach has been confirmed by multiple clinical trials, each reporting that light percepts, called phosphenes, are reliably elicited in response to electric stimulation from implanted electrodes. Unfortunately, the usefulness of vision elicited in this manner remains somewhat limited. For example, individual phosphenes do not reliably "assemble" into predictable spatial patterns, limiting the spatial information that can be conveyed to the patient. Even the description of individual phosphenes varies considerably with differences in size, color, shape and intensity all reported.
While many factors are likely to contribute to the limited quality of elicited vision, the use of sub-optimal methods of stimulation are thought to play a significant role. For example, electrode arrays are often positioned in close proximity to the nerve fiber layer—the layer in which ganglion cell axons course towards the optic disk. These axons are themselves highly sensitive to stimulus pulses, and therefore, distant ganglion cells whose axons pass closely to the stimulating electrode are likely to be activated. This distorts retinotopy (the correspondence between the spatial pattern of the stimulus and the spatial pattern of elicited neural activity) which is likely to contribute to the poor quality of percepts reported during clinical trials. A recent study from the lab suggests a method of robustly activating nearby neurons without simultaneously activating passing axons.
Nearly 15 years ago, Humayun, et al, showed that electric stimulation of the retina in blind subjects could elicit light percepts (phosphenes). Despite much effort however, the quality of elicited vision remains somewhat limited. To improve this, we are studying how and why retinal neurons respond to different forms of electric stimulation. By understanding the basic principles by which these neurons response to stimulation, we hope to be able to develop powerful new stimulation techniques that create specific patterns of neural activity—ideally, patterns that replicate one or more elements of the patterns used physiologically by the healthy retina.
Fundamental Mechanisms of Activation
We have mapped the threshold sensitivity of individual ganglion cells to electric stimulation from a small point-source electrode. We found threshold was lowest in and around the proximal axon of the cell. Immunochemical staining revealed that this region contained a dense "band" of voltage-gated sodium channels suggesting that the band is the site that is responsive to extracellular electric stimulation. Interestingly, we found that the length and location of bands were different in different types of ganglion cells. A computational model was developed to explore how these changes influence sensitivity to stimulation and revealed that band length and distance from the soma were both correlated with lower thresholds. We continue to explore the precise factors that influence activation using a combination of electrophysiology, immunochemical staining and computational studies.
Alternative Stimulation Waveforms
Pulsatile waveforms have been used widely to activate neurons in both the peripheral and central nervous system. The sharp temporal change in voltage is thought to be optimum for activating the voltage-gated sodium channels known to underlie the action potential. In the retina however, there are neurons that do not contain dense regions of sodium channels and yet these neurons still respond strongly to stimulation. This raises the possibility that other types of voltage-gated ion channels can also contribute to or even underlie activation. In a series of experiments, we found that low-frequency sinusoidal waveforms generated robust activity in the neurons that do not contain sodium channels but do not activate those neurons that do contain dense sodium channels. This is highly important because it suggests that such waveforms can be used to generate focal activity around each stimulating electrode without simultaneously activating the axons of passage from distal neurons. We continue to explore the mechanism underlying this non-sodium-channel activation—a recent modelling study confirms our hypothesis that voltage-gated calcium channels are a strong contributor.