The rapid discrimination of auditory location information enables grouping and selectively attending to specific sound sources. The typical indicator of auditory change detection is the mismatch negativity (MMN) occurring at a latency of about 100–250 ms. However, recent studies have revealed the existence of earlier markers of frequency deviance detection in the middle-latency response (MLR). Here, we measured the MLR and MMN to changes in sound location. Clicks were presented in either the left or right hemifields during oddball (rare 30°-shifts in location), reversed oddball, and control (sounds occurring equiprobably from five locations) conditions. Clicks at deviant locations elicited an MMN and an enhanced Na component of the MLR peaking at 20 ms compared to clicks at standard or control locations. Whereas MMN was not significantly lateralized, the Na effect showed a contralateral dominance. These findings indicate that, also for sound location changes, early detection processes exist upstream of MMN.
Spatial information provides a critical cue to separate auditory objects and plays a crucial role in auditory scene analysis. The extraction of location information enables us to selectively attend to a specific source of auditory input at a given moment in time and to monitor the auditory environment for the occurrence of new sound-emitting objects.
In human electrophysiology, the detection of contextually new or rare auditory features is usually associated with the mismatch negativity (MMN) of the event-related brain potential (ERP). The MMN is a component that is elicited by rare and irregular sounds (termed deviants) occurring in the context of an otherwise repetitive or regular stimulation, irrespectively of the subject's focus of attention (Escera, 2007 and Näätänen et al., 1978). Recent theoretical models on novelty and deviance detection and their relation to attention assume that the representation of the incoming stimulus is automatically matched to the predicted sensory information derived from previous stimulus regularities that are held in auditory sensory memory (e.g. Näätänen et al., 2011 and Winkler et al., 2009). Importantly, the matching process is supposed to happen on the level of an integrated stimulus representation, i.e. after feature extraction and integration have been completed resulting in the typical MMN latency between 100 and 250 ms. The neural generators of the MMN have been modeled in the supratemporal cortex of the two hemispheres (Alho, 1995 and Huotilainen et al., 1998) and, depending on the paradigm, additionally in frontal areas (Deouell, 2007). Intracranial recordings in animals suggest that MMN-like responses can be observed in different brain structures, including the primary and secondary auditory cortex (Csepe, 1995), the medial geniculate body of the thalamus (Kraus et al., 1994), the inferior colliculus (Csepe, 1995 and Malmierca et al., 2009), and also the hippocampus (Ruusuvirta et al., 1995). Based on the findings of single-unit recordings, a property of auditory neurons termed stimulus-specific adaptation (SSA; Antunes et al., 2010, Malmierca et al., 2009, Ulanovsky et al., 2003 and von der Behrens et al., 2009) has been proposed as a neural mechanism involved in deviant processing and the generation of MMN (Nelken and Ulanovsky, 2007). An important implication of those studies in animals is the hypothesis that the processing of contextually novel or rare sounds can act at multiple time-scales and different anatomical levels along the auditory pathway starting in the animal model already at about 15 ms after stimulus onset (Pérez-González et al., 2005 and von der Behrens et al., 2009). Recent human studies suggest that a similar cascade of deviance detection processes spanning over multiple time scales is at work in the human auditory system. This was shown by Grimm et al. (2011) who reported an enhancement of the Nb component (40 ms after sound onset) of the middle-latency response (MLR) in addition and prior to the elicitation of MMN for rare frequency deviants. A modulation of an MLR component (namely, the Pa component peaking at about 30 ms) through the deviance status of a sound was also reported by Slabu et al. (2010) for deviants differing in their spectral content using broadband-filtered noise stimuli.
