The ability to localize moving sound sources is critically important for our interaction with the acoustical environment. Auditory motion perception has been studied over many years in terms of localization accuracy and resolution (see for reviews Grantham, 1997; Carlile and Leung, 2016). In the auditory modality, spatial resolution or least perceptible difference in spatial location of stationary sources is estimated by the minimum audible angle (MAA). For locations near the head midline, the MAA ranges from 1 deg to 3 deg depending on sound frequency, and increases with more lateral locations (to 7 deg or more) (Mills, 1958; Perrot and Pacheco, 1989; Thavam and Dietz, 2018). Spatial resolution for moving sounds is represented by the minimum distance that a stimulus needs to travel to be distinguished from a stationary source. This just-noticeable angular distance is referred to as the minimum audible movement angle (MAMA). The MAMA rapidly increases for stimulus durations below 200 ms, and grows almost linearly from 8 deg at 90 deg/s to 21 deg at 360 deg/s. Moreover, the time needed to detect sound movement decreases with growing velocity (Perrott and Musicant, 1977; Grantham, 1986; Carlile and Leung, 2016).

Importantly, the MAMA for azimuthal sound motion is usually 2 to 3 times larger than the MAA, which is due to the fact that auditory system responds more slowly to changing localization cues than to constant ones (Blauert, 1972; Grantham and Wightman, 1978; Saberi and Hafter, 1997). The limited ability of the binaural system to follow ongoing changes in the interaural pattern of stimulation (interaural differences in time (ITD), phase (IPD) or level (ILD)) is called “binaural sluggishness”. This phenomenon is widely considered to be reflecting the integration time of the binaural system, that is, the time interval required for integrating the acoustical changes and for detecting the displacement of a sound source. This integration window defines the limits of temporal resolution of spatial hearing.

The time constants of binaural system may be estimated differently, depending on psychophysical tasks, procedures, and stimuli employed. The subjective sensation of motion may arise from sequential presentation of short sound bursts. In this context, the temporal boundary for perception of apparent motion is determined by the stimulus onset asynchrony (SOA), and the optimal SOA values fall within the range from 30 ms to 100 ms (Altman and Romanov, 1988; Strybel et al., 1990). The studies of dynamic spatial resolution suggested the concept of low pass filtering of interaural information, which implies averaging of the successive parts of the stimulus (Saberi and Hafter, 1997). The minimum integration time that is required for detecting changes in location or in velocity was estimated at between 100 to 1000 ms, which is several times longer than for localizing stationary sounds (Chandler and Grantham 1992; Saberi and Hafter, 1997; Carlile and Leung, 2016).

Temporal aspects of auditory processing of moving sounds can be efficiently studied using the event-related potentials (ERPs). In order to separate the response evoked by motion from the one evoked by sound onset, several studies employed the delayed-motion paradigm in which the onset of the sound motion is delayed relative to the sound onset (e.g., Krumbholz et al., 2007; Getzmann, 2009). In these conditions, a motion-specific response (motion onset response, MOR) can be identified after the stimulus onset response. A MOR typically consists of early negative deflection (“change”- N1, cN1) and late positive deflection (“change”- P2, cP2). MOR latency is longer than that of the sound onset response (at least 140 ms relative to the start of motion). The MOR differs in morphology and topography from the response to sound energy onset, and can be modulated by stimulus properties such as spatial motion cues, motion direction and velocity (Getzmann, 2009; Getzmann and Lewald, 2010).

Numerous studies have reported that MOR can be evoked by stimuli moving at the velocities of 60 deg/s and higher (Krumbholz et al., 2007; Getzmann, 2009; Getzmann and Lewald, 2010, 2012; Grzeschik et al., 2010, 2013), whereas behavioral measurements indicated above-threshold discrimination for much slower sound movement (Grantham, 1986; Harris and Sergeant, 1971; Sabery and Perrott, 1990; Carlile and Best, 2002; Altman and Viskov, 1977; Getzmann et al., 2004; Schmiedchen et al., 2013; Leung et al., 2016; Carlile and Leung, 2016 (a review)). The relationship between behavioral and electrophysiological temporal measures of sound motion processing has not been properly studied.

Meanwhile, behavioral correlates of brain activity were investigated in greater detail in visual perception. Patzwahl and Zanker (2000) varied the strength of visual motion in terms of motion coherence and recorded the ERPs evoked by onset of motion, in parallel with measuring reaction time (Patzwahl and Zanker, 2000). They reported that both reaction times and ERPs were influenced systematically by the strength of the motion signal. Importantly, the latency of motion-evoked response decreased linearly with increasing motion coherence. Later it was found that both the latency of motion-evoked response and reaction time decreased as motion velocity increased, and both measures were well approximated by the same negative power function with the exponent close to -2/3 (Kreegipuu and Allik, 2007).

Effect of auditory motion velocity on reaction time and MOR was first investigated by Getzmann (2009). In his reaction-time task, listeners responded as fast as possible to the motion onset. The results suggested that both MORs and reaction times were ordered according to velocity, in a way that higher motion velocities were associated with earlier and larger MORs as well as with shorter reaction times. Later, the reaction-time task was employed together with MOR recording to explore the effect of natural versus artificial spatial cues on cortical processing (Getzmann and Lewald, 2010). During the active trials, listeners performed a direction discrimination task. It was reported that free-field motion and virtual 3D-motion were associated with earlier cortical responses and with shorter reaction times than ILD- or ITD-motion. To sum up, these findings indicate a close correspondence of brain activity and reaction time in auditory motion detection at a group level.

Reaction time provides, however, merely an indirect perceptual estimate which comprises several stages of information processing, including response execution time. In search of a more precise estimate, we designed the psychophysical part of the current study in order to determine temporal thresholds for perception of motion at various velocities in the delayed-motion paradigm typically employed to obtain the MOR potential. The length of the moving sound segment progressively decreased along with the angular shift of the stimuli. Psychophysical procedure was determined by the task of measuring MAMAs. The minimal duration of motion, at which the listeners could identify its direction, was taken as a temporal measure. We anticipated that this time of direction identification should be inversely related to the motion velocity.

Our experiment aimed to find whether MOR latency could be approximated by a function of velocity similar to the direction identification time. If cN1 latency is inversely related to velocity, this would imply that the time during which the information about sound motion is accumulated by the MOR-generating system is related to the critical value of angular shift, the way it is in perception. Then the MOR latency can be viewed as a neurophysiological index of the temporal integration occurring during the perception of moving sound. If not, then the time required for accumulation of spatial information does not depend on the angular distance traveled by sound. So, the main goal of our study was to establish the relationship between MOR latency and behavioral measures of temporal integration. This relationship may be important in the context of neuronal mechanisms involved in the processing of any dynamic changes in the environment. The second question under study was the potential existence of an integration time window for motion processing during which a stimulus should move in order to evoke the MOR. We expected that this time window, determined by psychophysical measurements, could set the boundary conditions for the MOR produced by slow motion.