Rhythmic gain control during supramodal integration of approximate number

Highlights

• Deconvolution of ongoing EEG signals during sequential numerosity integration

• Endogenous oscillatory phase ~ 3 Hz predictive of subsequent numerosity judgment

• Supramodal rhythmic gating of visual, auditory, and somatosensory inputs

• Rhythmic gain control during decision-formation, but not during perceptual encoding

Abstract

According to one view, neural oscillations structure information processing in time, determining whether sensory inputs have a strong or weak impact on behavior. Recent work showed that during sequential integration of visual inputs, stimuli that fall in the preferred phase of slow (1–3 Hz), endogenous EEG activity carry greater weight in subsequent judgment. Here, we asked two questions. Firstly, is this phenomenon modality-specific, or is it supramodal? Secondly, does this effect occur at the level of sequential encoding, or only during decision formation? We analyzed scalp EEG recordings from healthy human participants while they compared the approximate number of visual, auditory or somatosensory pulses in two successive intervals (N1 and N2). Despite differences in activity evoked in different domains, a common, slowly-oscillating (~ 3 Hz) choice-predictive signal was observed in all three modalities with a maximum coincident with pulse onset. Critically, this signal was present during N2 (when a decision was being formed) but absent during N1 (when perceptual information was encoded, but no decision could be made). In other words, rhythmic gain control during sequential processing is a supramodal phenomenon that occurs while information is integrated towards a categorical decision.

Introduction

The electrical activity produced by the central nervous system tends to exhibit oscillatory dynamics, both at the level of local microcircuits and across large-scale cortical networks. A longstanding controversy concerns the role of neural oscillations in shaping the computations that underlie behavior. One view holds that electrical oscillations are epiphenomena of the biophysical mechanisms underlying neural circuits, and play no causal role in information processing or behavior (Shadlen and Movshon, 1999). An alternative account proposes that the phase, amplitude and frequency of neural oscillations actively encode information relevant to a current task (Buzsáki and Draguhn, 2004). One emerging theory suggests that during perception, neural oscillations might act to structure information processing across time, by periodically gating neural excitability and thereby providing temporal windows of privileged information processing and transfer (Schroeder and Lakatos, 2009, VanRullen et al., 2011). For example, in the sensory cortices of macaque monkeys, the gain of single-neuron responses is heightened when they fall in phase with oscillations entrained by rhythmic stimulation (Lakatos et al., 2007, Lakatos et al., 2008). Consistently, rhythmic entrainment of cortical oscillations was found to modulate human observers' performance in various perceptual tasks (e.g., Cravo et al., 2013, Henry and Obleser, 2012, Ng et al., 2012, Stefanics et al., 2010).

Evidence is mounting that psychophysical performance also depends on the phase of spontaneous, endogenously ongoing brain rhythms. Visual, auditory, or tactile stimuli are more likely to be detected or discriminated when they fall in the preferred phase of ongoing cortical oscillations near the alpha-frequency range (Ai and Ro, 2014, Busch et al., 2009, Hanslmayr et al., 2013, Mathewson et al., 2009, Strauß et al., 2015). Complementing these findings, endogenous rhythmic fluctuations at a lower frequency (delta, ~ 1–3 Hz) were recently reported during sequential accumulation of multiple inputs over time (Wyart et al., 2012). In that study, observers integrated the information provided by successive visual events occurring in rapid sequence to make a category judgment. Those events that fell in the preferred phase of an endogenous delta oscillation were found to be more influential in determining choices, as if momentary inputs were rhythmically gated during entry to a cumulative decision variable. Low-frequency oscillatory gating of sequential information may contribute to attentional selection and temporal anticipation (Cravo et al., 2013, Stefanics et al., 2010), and may ultimately provide periodic refractory periods that guard against catastrophic interference and information overload in neural processing (Marois and Ivanoff, 2005, Sergent et al., 2005).

However, although support for this view is growing, two factors limit the conclusions that can be drawn from existing studies. Firstly, slow endogenous fluctuations in sequential processing have thus far been studied mostly in the visual domain (Wyart et al., 2012, Wyart et al., 2015; but see e.g., Giraud and Poeppel, 2012 for discussion of low-frequency oscillations in speech processing). It is thus unclear whether phase-dependent gain control of to-be-accumulated sensory inputs is a ubiquitous, supramodal phenomenon, or one that depends on the modality of sensory input. Secondly, the processing stage at which gating occurs remains unclear. Signals could be gated at an early processing stage, during sequential encoding, or later, during conversion to a decision signal, and extant studies have failed to distinguish between these possibilities.

Here, we addressed these two questions directly. To do this, we capitalized on data from a previously published experiment in which human observers compared the approximate number of pulses occurring in two successive intervals (N1 and N2; see Spitzer et al., 2014). In separate conditions, pulses were administered in the visual (blinks) auditory (beeps), and somatosensory (electrical pulses) domains. Here, we used convolution modeling of scalp electroencephalographic (EEG) data to estimate how the weight (or choice-predictiveness) associated with each pulse varied over time. We observed evidence for a rhythmic, phase-dependent gain control in centro-parietal signals at ~ 3 Hz in all three modalities, suggestive of a supramodal phenomenon. Critically, this modulation accompanied decision formation in interval N2 but was entirely absent in interval N1, when participants accumulated the pulses in the reference interval for later comparison. In other words, endogenous, delta-rhythmic gain control during perceptual decisions occurs at a late processing stage, during conversion of perceptual signals into a decision variable.