The perceptual categorization of multidimensional stimuli is hierarchically organized

Published: 3 February 2023| Version 1 | DOI: 10.17632/jgsv9xgwk8.1
CHi Chen,
Livia de Hoz


As we interact with our surroundings, we encounter the same or similar objects from different perspectives and are compelled to generalize. For example, we recognize dog barks as a distinct class of sound, despite the variety of individual barks. While we have some understanding of how generalization is done along a single stimulus dimension, such as frequency or color, natural stimuli are identifiable by a combination of dimensions. To understand perception, measuring the interaction across stimulus dimensions is essential. For example, when identifying a sound, does our brain focus on a specific dimension or a combination, such as its frequency and duration? Furthermore, does the relative relevance of each dimension reflect its contribution to the natural sensory environment? Using a 2-dimension discrimination task for mice we tested untrained generalization across several pairs of auditory dimensions. We uncovered a perceptual hierarchy over the tested dimensions that was dominated by the sound’s spectral composition.


Steps to reproduce

Animals 5-6 Weeks Female C57BL/6JOlaHsd mice were used. Apparatus The Audiobox was a device developed for auditory research from the Intellicage (TSE, Germany). The Audiobox served both as living quarters for the mice and as their testing arena. The Audiobox consisted of three parts: a home cage, a drinking ‘chamber’, and a long connecting corridor. The presence of the mouse in the ‘chamber’ is a ‘visit’. Sounds Mice were trained with pairs of either frequency-modulated (FM) sweeps, amplitude-modulation (AM) sounds, or pure tone pips. For 2-d discrimination task, sound pairs used in training as safe and conditioned differed in two out of four chosen dimensions. For the FM sweep sounds these dimensions were frequency range, duration, sweep direction of modulation and velocity of the sweep. Frequency sweeps had a duration of 20 ms (default) or 40 ms, including 5 ms rise/decay, and one of four modulation velocities (50, 62.5, 87.5 or 100 octave/sec; with 50 octaves/s as default). For the AM stimuli, tested dimensions were carrier frequency and modulation rate. The AM sounds had 100% sinusoidal modulation and had one of four carrier frequencies (6670, 8404, 10588 or 13340 Hz), as well as one of four modulation rates (5, 8, 12 or 20 Hz). For pure tone pips, the dimension of tone frequency and repetition rate were tested. Similar to the AM stimuli, the pure tones pips had one of four frequencies (6670, 8404, 10588 or 13340 Hz) and one of four repetition rates (2, 3, 5 or 8 Hz). The frequency-only discrimination: mice trained with upward FMs differed in frequency range (3 to 6 kHz vs. 9 to 18 kHz). For direction discrimination, the mice were trained to discriminate between sweeps of opposite direction of modulation. For duration discrimination, mice were trained to discriminate FM upward sweeps that differed only in duration (20ms vs. 80ms). Discrimination task Mice were trained to discriminate two sounds, the ‘safe’ and ‘unsafe sound, in the Audiobox. One sound was always ‘safe’, meaning that access to water during these visits was granted upon nose-poke. For the first 5 days, only the safe tone was played. Another sound, a ‘unsafe one, was introduced in ‘unsafe' visits and a nose-poke during these visits was associated with an air puff. The probability of unsafe visits was 9.1% for the first 2 days, increased to 16.7% for the next 2 days, then stayed at 28.6% until they showed steady discrimination performance for at least 3 consecutive days. Generalization measurement During generalization testing, visits consisted of 55.6% of safe visits, 22.2% of conditioned visits and 22.2% of novel visits in which a novel sound was presented. Nose poking during the presentation of the novel tone resulted in opening of the doors that gave access to water. Novel sounds represented all possible combinations of values used along each dimension. On each testing day, only two novel sounds were presented (in 11.1 % of total visits each).


Neuroscience, Animal Behavior, Auditory System, Cognitive Neuroscience