We study the sense of touch at different levels of analysis: anatomy, physiology, psychophysics, behavior, computational modeling, and applications. By using standard light microscopy and fluorescence microscopy, we visualize the structure of skin, distribution of nerve endings and the mechanoreceptors. We use standard Nissl stain and additional dyes to study cytoarchitecture, electrode tracks, microinjection sites in the somatosensory cortex. While stimulating the skin surface mechanically, we record single-unit spikes from peripheral nerves to understand the response properties of mechanoreceptive fibers in various conditions. For example, spike activity is measured during the application of pharmacological agents on the skin. We also record spikes from the somatosensory cortex in acute preparations and study the transformation of tactile information and the effects of drugs applied by microinjection. Anatomical and neurophysiological experiments are run on animal subjects. By using similar stimulus protocols, we perform psychophysical experiments on humans. Data collected from animals are used to establish computational models which are subsequently tested against the results of human psychophysical experiments. Recently, we have also been training rats, implanted with chronic electrodes, in various behavioral tasks to develop a sensory neuroprosthesis.


Anatomy: Skin is a complex organ and contains all tissue types: epithelial, connective, muscle, and nervous tissue. Although elaborate skin models exist in the literature regarding mechanical properties, we prefer simpler models which can be combined with the neurophysiological models to predict the sensory responses measured in our laboratory. For this purpose, we obtain morphometric measures on skin structure, nerve endings and mechanoreceptors. There are four types of mechanoreceptors in the mammalian glabrous skin: Meissner corpuscles, Pacinian corpuscles, Merkel cell-neurite complexes, and Ruffini endings. The distribution of these mechanoreceptors and the innervation pattern of their afferent fibers vary significantly. By immunohistochemical methods, we localize these structures within the skin and quantify their densities and distributions. We also look for specific proteins in the neural and accessory structures of the mechanoreceptors in order to find out about their functional roles in tactile sensation. For example, structures related with the chemical transmission in the Merkel cell-neurite complex is one of our current concentration areas. Skin samples are typically collected from frogs, rats, and rabbits. We also perform routine Nissl staining on rat brain sections. We specifically study the hindpaw representation in the primary somatosensory cortex and identify recording/injection sites used in the physiology experiments.


Physiology: After routine classification of mechanoreceptive fibers dissected from peripheral nerves, we apply mechanical vibrations on the skin surface and record single-unit spikes. In addition to standard spike histograms, we use advanced mathematical techniques to study response properties as the stimulus parameters change. For pharmacological manipulations, we perfuse skin patches in a tissue bath and study drug effects on the response properties. This helps us to understand mechanoelectric transduction in the receptors. For example, our recent results confirm fast serotonin transmission in the Merkel cell-neurite complex. Neurophysiological results are used to simulate the population response of afferents which project to the central nervous system. Our current work involves recording from tactile neurons in the somatosensory cortex. We use microelectrodes under acute stereotactic surgery to record single-unit spikes. Our focus is the transformation of the afferent population response to representations in the cortex. We are particularly interested in the excitation-inhibition balance and neural codes which shape psychophysical responses. A considerable computational effort is required to make sense of the spike data recorded from the cortex and to incorporate them to reliable computational models. Peripheral recordings are done in frogs and rats; cortical recordings are done in rats. We have ongoing microinjection studies on rat cortex, and study the attentional modulation of tactile signals.


Psychophysics: The sense of touch arises from the fusion of four psychophysical channels corresponding to four receptor systems. We study how tactile information is encoded in these channels individually, and how information is integrated to form the percept. By using stimulation protocols similar to those used in our neurophysiological experiments, we perform psychophysical experiments on human subjects. Some of these are complemented by EEG or evoked potential recordings from the scalp. In addition to objective measures in detection and discrimination tasks, we also collect subjective estimation and matching data. The results of psychophysical experiments are compared to predictions of computational models based on animal data. For example, we had shown that vibrotactile thresholds are approximately constant when stimulating successive distal locations in the finger, although innervation density increases towards the fingertip. Preliminary neurophysiological results explain this discrepancy based on spatial parameters of the stimulus and its associated representation in the afferent population response. Another focus of our work is the neural basis of psychophysical forward masking and temporal summation. Our laboratory also studies higher-order processes in cognitive tasks involving active touch and mental rotation of tactile objects.


Computational models: Models are developed to establish an overall view of tactile processing. They help us to formulate new hypotheses and to guide future research. As the models improve, they trigger engineering applications. We mostly use MATLAB for our computational analyses. Fundamental knowledge of basic physical and life sciences, and related engineering topics is essential in modeling. We mostly utilize modeling approaches used in electrical and mechanical systems. Therefore, mathematical techniques used in linear and nonlinear system analysis, circuit and network theory, signal processing, signal detection theory, filtering and time-series analysis, time-frequency analysis, probability and statistics, random processes, information theory, optimization, multivariate analysis, finite elements, etc. are helpful tools to model neural processing in the sense of touch.


Applications: Most of our basic research can be applied to certain clinical problems, especially in neurology and psychiatry, which involve somatosensory dysfunction. The sense of touch may be disrupted due to trauma, stroke, infection, metabolic (e.g. diabetic neuropathy) and autoimmune (e.g. Guillain-Barré syndrome) diseases, or psychiatric conditions which affect sensory processing (e.g. autism). Accurate understanding of function in the normal state is a prerequisite for medical diagnosis and treatment. Our laboratory is currently studying tactile processing in autistic children and our pharmacological work on sensory transduction may be clinically relevant in neuropathies. Additionally, our studies on skin morphology and mechanics may have implications in dermatology and plastic surgery. An interesting offshoot from our work will be the development of neural prosthetics. We are particularly interested in sensory substitution and augmentation. We have developed a special vibrotactile system to be used in an operant chamber for rats. We train rats, which are chronically implanted with multielectrodes, for various tactile tasks. Our preliminary results are promising that intracortical microstimulation of rats may substitute for vibrotactile stimulation of the glabrous skin. These efforts will be useful not only for humans, but also for robotic applications. Recently, we developed a prototype tactile sensor for the automatic palpation of breast tumors. Psychophysical testing is helpful for setting the limits of sensory processing in prosthetics and artificial devices. We are interested in how blind people process tactile information. The results of our cognitive experiments may be useful for designing tactile aids. Ideas generated in the Tactile Research Laboratory can improve products in the consumer market as well. People touch paper products, cleaning agents, textiles, food products, etc. continuously in their daily lives. We have previously signed a contract for the tactile surface design of beverage containers.


Institute of Biomedical Engineering.







From Press et al. (2010)


From Ayyildiz et al. (2013)