Importance of spike timing in touch: an analogy with hearing?

Abstract

Touch is often conceived as a spatial sense akin to vision. However, touch also involves the transduction and processing of signals that vary rapidly over time, inviting comparisons with hearing. In both sensory systems, first order afferents produce spiking responses that are temporally precise and the timing of their responses carries stimulus information. The precision and informativeness of spike timing in the two systems invites the possibility that both implement similar mechanisms to extract behaviorally relevant information from these precisely timed responses. Here, we explore the putative roles of spike timing in touch and hearing and discuss common mechanisms that may be involved in processing temporal spiking patterns.

Figure 1. Exploiting first spike latencies in hearing and touch

A∣ Precise spike timing is used in hearing to localize sound sources. Sound from a source towards the left will excite hair cells in the left ear (L) before hair cells in the right ear (R). Precisely timed excitatory and inhibitory inputs will reach an output cell (O) at different times, determining the strength of the response. B∣ Potential use of delay lines in touch. Touching an object of a given curvature will excite some tactile afferents earlier than others. Nerve fibers can be exploited as delay lines to detect the specific sequence of afferent firing by neurons in cuneate nucleus.

Figure 2. Coding of skin vibrations in the somatosensory nerves

A∣ Absolute threshold for three types of cutaneous mechanoreceptive afferents over a range of frequencies. Different afferent types are most sensitive at different frequencies. B∣ Neural responses (red) of a PC afferent to five repeated presentations of a 400 Hz sinusoidal skin oscillation (black). Responses are repeatable and tightly locked to the stimulus waveform. C∣ Responses of a PC afferent to 30 presentations of a texture (nylon, see inset showing the surface profile of a 7×7 mm patch), scanned across the skin at 80 mm/s. The fine structure of the texture is reflected in precise and repeatable temporal spiking patterns.

Figure 3. Potential mechanisms for exploiting precise spike timing in pitch processing

A∣ Delay lines can be used to detect periodic spiking activity in the nerve at a given frequency. In this illustration, I denotes the input neuron, while O denotes the output neuron, which receives input from I through a fast connection (1) as well as a delay line (2), which temporally shifts the neural responses by a constant delay. Coincidence detection will confer to O a selectivity for a frequency determined by the delay. B∣ Coincidence detection on several inputs can be used in order to extract the fundamental frequency from complex harmonic tones. Here, the output neuron receives input from two neurons. Coincident spikes in the input (indicated in red) elicit output spikes, while other spikes will not. C∣ Lateral inhibition mediated by detecting phase differences in neighboring neurons can be used to sharpen spatial representations of spectral frequency composition. Inhibitory interneurons are denoted by L. The spatially distributed response at the periphery results in a broad spatial pattern of activation which is converted to a sparser spatial representation through precisely timed inhibition.

Figure 4. Auditory timbre and tactile texture

A∣ Spectrograms of two tones (F3 and C3, different rows), played by two different instruments (horn and viola, different columns). B∣ Spectrograms of skin oscillations elicited by two different textures (vinyl and silk) when scanned at different speeds across the fingertip skin (80 and 120 mm/s). C,D∣ Time-averaged power spectra of sound waves (C) and skin vibrations (D) when correcting for fundamental frequency/scanning speed by shifting the spectra of the lower tone/speed (red trace) to the higher one (black trace). The general spectral composition is preserved across pitch/speed and can be used to resolve different instruments or textures. [Sound files were obtained from the University of Iowa Electronic Music Studios database, http://theremin.music.uiowa.edu]