While we think that much of the sound information is distributed globally,

it is remarkable that a good prediction of the categorization behavior and of its variability could even be obtained with well-chosen single local populations (Figure 8). This indicates that the perceptual decisions made by the animal can be parsimoniously explained by the selection of a relatively small group of neurons that spontaneously provides a suitable categorization of sound stimuli. Such a model would be Autophagy inhibitor in vivo distinct from scenarios in which learning leads to an optimal adaptation of a plastic decision boundary to a continuous sensory representation. What may be the functional role of discrete dynamics in circuits of the auditory cortex? In the primate visual system, complex category signals (objects or groups of

objects) are classically reported in higher areas such as inferior temporal, parietal, and prefrontal areas (DiCarlo et al., 2012; Swaminathan and Freedman, buy GW-572016 2012). The observation of category-forming dynamics already in a primary sensory area suggests that this may be a general property of neocortical circuits. It is conceivable that higher-order categories are built on a hierarchy of lower-order categories which arise in primary sensory areas. Such a hierarchical structure of discrete representations might be essential for elaborate cognitive functions such as language processing. The fact that, e.g., phonemes are perceived and thereby stably recognized as discrete sound categories (Liberman et al., 1967) might rely on similar dynamics of the human auditory cortex. Experimental subjects were male CB57BL/6J mice (Charles River, age: 6 to 16 weeks) and were

performed in accordance with the Austrian laboratory animal law guidelines and approved by the Viennese Magistratsabteilung MA58 (Approval #: M58/02182/2007/11; M58/02063/2008/8). All sounds were delivered free field at 192 kHz sampling rate in a sound proof booth by a custom-made system consisting of a linear amplifier and a ribbon loudspeaker placed nearly 25 cm from the mouse head (Audiocomm, Vienna, Austria). The transfer function between the loudspeaker and the location of the mouse hear was measured using a probe microphone (4939-L-002, Brüel&Kjær, Bremen, Germany) and compensated numerically by filtering the sound files with the inverse transfer function to obtain a flat frequency response at the mouse ear (between 0.5 kHz and 64 kHz ± 4 dB). Sound control and equalization was performed by a custom Matlab program running on a standard personal computer equipped with a Lynx 22 sound card (Lynx Studio Technology, Inc, Costa Mesa, CA).