Active perception of virtual texture frequency in the whisker-related sensorimotor system of the rat

Abstract

Rats are able to sample their surroundings for tactile discrimination purposes by performing series of whisker deflections. In this perceptual state, called active perception, the animal chooses how to best manipulate its mobile sensors (the whiskers) based on the incoming sensory information it receives via these sensors, so as to perform better at the task at hand. On the other end of the perceptual spectrum lies passive perception which can be described as a “wake-up” call to investigate the surroundings if a sensation was received which was not expected by the subject. During active perception each deflection leads to a mechanical stress at the base of the whisker. Additionally, the touch of any object which the whisker finds along its path also leads to whisker deflections and hence whisker vibrations. These combined vibrations are termed the vibrotactile signal. The nature of the vibrotactile signal is highly complex and it is still not known which of its physical parameters are used by the animal for discrimination. Moreover, it is not yet clear if the tactile system of the rat integrates over this signal to reduce its complexity or if the animals have access to instantaneous kinematic parameters such as minute details of the whisker trajectory. Some of the possible parameters which the animals could use for performing tactile discrimination include the temporal frequency of the signal (i.e. its spectral information), its intensity (i.e. the mean speed of the signal) or instantaneous kinematic features (such as details of the whisker trajectory). In this work I established a new psychophysical Go/No-Go paradigm for spatial frequency discrimination using actively whisking head fixed rats in a virtual environment where a natural stimulus is replaced by electrical microstimulation of the ascending sensory pathway. The major advantage of replacing natural stimuli such as textures with electrical stimulation is the greater level of stimulus control achieved from trial to trial. The initial goal of the project was to investigate the differences in neuronal activation in the barrel and motor cortex arising in active versus passive perception. For the purpose I wanted to train animals to discriminate sets of virtual grids with defined line spacing first in an active case- where the animal has to sweep its whisker in space and receive an electrical pulse in the primary somatosensory cortex each time a grid line is crossed, and then in a passive case- in which the same animal is retrained to keep its whiskers still and the stimulation patterns from the active case are replayed in the cortex. To my great surprise, even after extensive training none of the experimental animals was able to discriminate between a set of two virtual grids with significantly different spatial frequencies. Thus I was unable to proceed towards my initial goal and show that the motor program of the animal would be changed in order to optimize sensory percept in active versus passive case. Instead, I provided the animals with an additional cue, increasing the amplitude of the electrical stimulation on a No-Go trial as compared to a Go trial, which led to immediate discrimination for all tested animals. This finding speaks against pure frequency discrimination but rather indicates that rats are more able to use instantaneous events such as the electrical pulse amplitude cues. This new knowledge implied that the current working hypothesis that frequency is the most important parameter for rats during active or passive perception had to be revisited. I thus abandoned the pursuit to find the differences between active and passive perception, and concentrated on fleshing out if the instantaneous kinematic events as given by the electrical pulse amplitude were the actual parameter of a surface which the animals use for active discrimination. To thoroughly investigate this question, I varied several parameters of the virtual grids used in the active whisking task described here, including their spatial frequency, the electrical pulse amplitude and the position of the grid along the whisking cycle. I thus investigated an additional set of stimuli in which the electrical microstimulation amplitude in a Go trial was increased in even steps until it reached the value set for a No-Go stimulus. The stimuli providing an electrical pulse amplitude cue were easily discriminated by the animals whereas the stimuli differing only in spatial frequency were more challenging for the animals, leading to an increased amount of false responses. An experiment in which the virtual grids varied only in spatial frequency and starting position along the whisking cycle, thus blurring position information, proved the behaviorally most difficult one for the animals. The performance improved in a control experiment which kept the virtual grid arrangement same as in this experiment but provided the No-Go grids with electrical pulse amplitude cue. These results show that the rats were able to use the position information given by the virtual grids for discrimination. Altogether, one can conclude that the rats were not using the temporal frequency cues but used the electrical amplitude and grid position cue when available.