Neural activity before and after conscious perception in dichotic listening

Abstract

The neural basis of conscious perception can be studied using stimuli that elicit different percepts on different occasions (multistable perception). Multistable perception allows direct comparisons between brain activity and conscious perception that control for sensory input, and also serves as a model for attentional competition, with the winning perceptual outcome varying across trials. Dichotic listening tasks present multistable stimuli consisting of two different consonant-vowels (CVs, one/ear). For each trial one ear usually conveys the dominant percept. We used EEG to measure neural activity before and after dichotic stimulus presentation to compare activity among left vs. right ear percepts and a control task. Consonant-vowels were perceived more often to the right vs. left ear. Pre-stimulus EEG power in the beta band (16–20 Hz) increased for left compared to right ear percepts and control trials. Event-related potentials after stimulus onset showed smaller P50 amplitudes (∼50 ms latency) for left ear compared to right ear and control trials. Results indicate that neural activity for right ear percepts is comparable to control conditions, while activity for the atypical left ear percept differs before and after stimulus onset. Pre-stimulus EEG changes for left ear percepts may indicate a mechanism of spontaneous fluctuations in cortical networks that bias attentional competition during subsequent sensory processing. The P50 amplitude differences among perceived ears suggests that rapid sensory and/or arousal-related activities contribute to the content of conscious perception, possibly by biasing attentional competition away from the dominant right ear channel.

Introduction

Conscious perception is thought to result from the interplay of bottom-up sensory information and top-down influences of current brain state (Sterzer, Kleinschmidt, & Rees, 2009). Here the term “bottom-up” refers to information conveyed by sensory pathways that reliably indicates features of a stimulus, such as a sound's frequency or intensity. In this context, “top-down” factors refer to other brain processes that are not an obligatory response to a given set of stimulus features but can nonetheless influence perception. Evidence for the importance of top-down factors is provided by stimuli that can elicit different percepts on separate occasions (termed a “multistable” stimulus) (Leopold & Logothetis, 1999). Multistable perception has been an important topic since the era of Gestalt psychology, and prominent examples of the phenomenon include figure-ground reversals when viewing the 3-D Necker cube drawing or the Rubin face/vase image (Rock & Palmer, 1990).

Perception of multistable stimuli in the visual modality has been studied intensively using binocular rivalry. In binocular rivalry two different stimuli are simultaneously presented, one to each eye, for the duration of a trial. Subjects usually perceive only one of the stimuli at a time, but the subject's perception alternates irregularly between the stimulus presented to either the left or right eye every few seconds (Levelt, 1966, Tong et al., 2006, Walker, 1978). Studies using electroencephalography (EEG) have shown that perceptual shifts in binocular rivalry are associated with changes in continuous EEG and visual evoked responses (Cobb and Morton, 1967, Lansing, 1964, Srinivasan, 2004). Thus, neural activity associated with a change in the perceived eye may be attributed to a change in perception due to how the invariant sensory input is processed.

In the auditory modality perception of multistable stimuli can be studied using the dichotic listening paradigm. In dichotic listening two consonant-vowels (CVs) are presented simultaneously, one CV to each ear, and one CV is usually the dominant percept on a given trial (Broadbent, 1954, Hugdahl, 2003, Kimura, 1961, Kimura, 1967). Perception varies between trials such that the same CV pair can yield a dominant left or right ear percept (termed “bistable perception”). Although the dominant ear percept varies across trials, most subjects identify the right ear CV ∼50% more often than the left (Hugdahl, 2003, Kimura, 1961). This right ear advantage has been related to a contralateral bias in ascending inputs from ear to cortex (Kimura, 1967, Milner et al., 1968), possibly leading to right ear inputs having a stronger influence on left auditory cortex and associated areas vital for language processing (Fitch, Miller, & Tallal, 1997). Earlier studies have identified correlates of the right ear advantage to dichotic and diotic CVs when comparing across subjects having left or right ear advantages (Ahonniska et al., 1993). There is also a right ear advantage for brainstem responses to clicks (Sininger & Cone-Wesson, 2006), and a left ear advantage to tones for cortical potentials (Ofek & Pratt, 2004). The present study builds on previous work by examining EEG/ERP differences in left vs. right ear percepts within subjects, which provides a direct comparison of the neural correlates of left vs. right ear percepts.

The specific relations between top-down and bottom-up processing in dichotic listening is an unresolved issue. Structural models based on contralateral sensory inputs imply that perception in dichotic listening is governed by bottom-up processing. However, other influences besides neural connectivity, such as attention, are needed to explain why right ear CVs are not perceived on all trials (Broadbent, 1954). Indeed, the proportion of right vs. left ear percepts can be modified by instructing subjects before the stimuli are presented to focus their attention to either left or right ear (Hiscock & Stewart, 1984). Although attention has a robust influence on dichotic listening performance, the timing of attentional influences relative to stimulus processing (pre vs. post) has not been defined. The possibility that neural activity before the dichotic stimulus may influence the propensity to perceive left vs. right ear CVs is also supported by single-unit recording and neuroimaging studies (Arieli et al., 1996, Liang et al., 2002).

The present study used EEG to identify cortical mechanisms of bistable perception by quantifying electrical brain activity associated with left vs. right ear percepts of dichotic stimuli. Spontaneous, perhaps top-down, differences in pre-stimulus electrical activity were analyzed as a function of the subsequently perceived ear to test the structural model's prediction of no pre-stimulus differences between ear percepts. Post-stimulus analysis of event-related potentials (ERPs) tested whether neural activity evoked by dichotic stimuli varied between left and right ear percepts. The control task in this study used the same dichotic stimuli as in the dichotic listening task to control for the influence of stimulus features on ERP responses. Additionally, the control task provided a metric for assessing whether neural processing reflected a “right ear advantage” or a “left ear disadvantage,” which has not been resolved in previous behavioral studies.