Enhanced deviant responses in patterned relative to random sound sequences

Abstract

The brain draws on knowledge of statistical structure in the environment to facilitate detection of new events. Understanding the nature of this representation is a key challenge in sensory neuroscience. Specifically, it is unknown whether real-time perception of rapidly-unfolding sensory signals is driven by a coarse or detailed representation of the proximal stimulus history. We recorded electroencephalography brain responses to frequency outliers in regularly-patterned (REG) versus random (RAND) tone-pip sequences which were generated anew on each trial. REG and RAND sequences were matched in frequency content and span, only differing in the specific order of the tone-pips. Stimuli were very rapid, limiting conscious reasoning in favour of automatic processing of regularity. Listeners were naïve and performed an incidental visual task. Outliers within REG evoked a larger response than matched outliers in RAND. These effects arose rapidly (within 80 msec) and were underpinned by distinct sources from those classically associated with frequency-based deviance detection. These findings are consistent with the notion that the brain continually maintains a detailed representation of ongoing sensory input and that this representation shapes the processing of incoming information. Predominantly auditory-cortical sources code for frequency deviance whilst frontal sources are associated with tracking more complex sequence structure.

Keywords: Mismatch negativity; Orbitofrontal cortex; Prediction error; Predictive coding; Surprise.

Fig. 1

Stimuli and Behavioural responses. a: Procedure for selecting frequencies used for each stimulus. From the pool of 26; 13 adjacent values were chosen at random as candidate sequence frequencies (purple); 10 were selected for the sequence. Of the remaining tones; all except the frequencies closest to the sequence could potentially be outliers (orange); and from these a single value was chosen at random to be the outlier on that trial. b: Example set of stimuli for the four conditions; these were generated together from the same frequencies in order to match acoustic properties. c: Results from the behavioural experiment. Left: reaction times to outlier tones. Right: sensitivity (d’) to outlier tones. **p < .01, ***p < .001.

Fig. 2

Deviance-evoked responses. a: Time-domain response averaged over the 39 central channels which showed a significant deviance response. Shading shows the standard error of the mean over subjects. The three deflections in the response correspond to the three clusters shown in (b). b: three time-channel clusters showing a main effect of deviance; i.e., (REGo – REGno) + (RANDo – RANDno) c: Topography of the three main-effect ROIs; averaged over the temporal extent of each ROI. d: Topography of the effect of regularity: expressed by the contrast (REGo – REGno) – (RANDo – RANDno). Channels included in the statistical analysis are shown in black [these are the significant channels in (c)]. Channels showing an effect of deviance at any point during the cluster are highlighted in white. The average magnitude of the response to REGo and RANDo, within each ROI, is shown in the bar plots below. e,f: Source-level activity shown on a template cortical sheet. T-statistic maps thresholded at T = 2. All show average source activity taken over a time-window defined by ROI1 (80–145 msec) and ROI2 (165–245 msec) e: Main effect of deviance in ROI1 (top) and ROI2 (bottom). f: Effect of regularity on the deviance response in ROI1 (top) and ROI2 msec (bottom). Peak T-statistic are indicated. Abbreviations: STG - superior temporal gyrus, IPL - inferior parietal lobule, FG - fusiform gyrus, S/MFG - superior/middle frontal gyrus, TP - temporal pole, IPS - intraparietal sulcus, CS - central sulcus, OG - orbital gyrus, STS - superior temporal sulcus, MTG - middle temporal gyrus.

Fig. 3

Sequence-evoked responses. a. Sequence-evoked response. Shown in the main plot is the root-mean-square (RMS) of the signal over all channels, representing global field power; shading shows the standard error of the mean over subjects. Time period showing significant difference between REG and RAND conditions is indicated by a grey bar. Polarity-resolved topographies (across all channels) are shown for the onset response from 50 to 80 msec (inset; left) and the sustained response (700–3000 msec) to REG (inset; top) and RAND (inset; bottom). b: Offset response. Top: Evoked response averaged over 58 central channels showing an effect of regularity. Bottom: Topography of the response during the two time-windows covering significant clusters for the contrast (REG – RAND); channels showing an effect of deviance at any point during the cluster are highlighted in white.