Characterizing the roles of alpha and theta oscillations in multisensory attention

Abstract

Cortical alpha oscillations (8-13Hz) appear to play a role in suppressing distractions when just one sensory modality is being attended, but do they also contribute when attention is distributed over multiple sensory modalities? For an answer, we examined cortical oscillations in human subjects who were dividing attention between auditory and visual sequences. In Experiment 1, subjects performed an oddball task with auditory, visual, or simultaneous audiovisual sequences in separate blocks, while the electroencephalogram was recorded using high-density scalp electrodes. Alpha oscillations were present continuously over posterior regions while subjects were attending to auditory sequences. This supports the idea that the brain suppresses processing of visual input in order to advantage auditory processing. During a divided-attention audiovisual condition, an oddball (a rare, unusual stimulus) occurred in either the auditory or the visual domain, requiring that attention be divided between the two modalities. Fronto-central theta band (4-7Hz) activity was strongest in this audiovisual condition, when subjects monitored auditory and visual sequences simultaneously. Theta oscillations have been associated with both attention and with short-term memory. Experiment 2 sought to distinguish these possible roles of fronto-central theta activity during multisensory divided attention. Using a modified version of the oddball task from Experiment 1, Experiment 2 showed that differences in theta power among conditions were independent of short-term memory load. Ruling out theta's association with short-term memory, we conclude that fronto-central theta activity is likely a marker of multisensory divided attention.

Keywords: Alpha; Attention; EEG; Multisensory; Oscillations; Theta.

Figure 1

Diagram depicting visual and auditory sequences in which oddball stimuli are embedded. Note that luminance-levels and pitches in this diagram are meant for illustrative purposes only. For graphical purposes, changes in loudness are represented here by changes in brightness.

Figure 2

Topographies of log-transformed alpha-band (8–13 Hz) power (left) and theta-band (4–7 Hz) power (right) for each of the three conditions (top: Auditory, middle: Visual, bottom: Audio-Visual) during the first second of the stimulus presentation (0–1000 ms). Black circles mark electrode clusters of interest.

Figure 3

Top Row: Time-frequency transforms of the posterior electrode cluster for each of the three conditions (Auditory, Visual and Audio-Visual). Dotted white line represents stimulus onset; solid white line represents one second after stimulus onset. Horizontal black lines bracket the alpha frequency band (8–13 Hz). Bottom Row: Time-frequency transforms of the anterior electrode cluster for each of the three conditions (Auditory, Visual and Audio-Visual). Dotted white line represents stimulus onset; solid white line represents one second after stimulus onset. Horizontal black lines represent the theta frequency band analyzed (4–7 Hz).

Figure 4

Diagram depicting visual and auditory sequences in which oddball stimuli are embedded for each of the six conditions in Experiment 2. Note that luminance-levels and pitches in this diagram are meant for illustrative purposes only. Each panel shows an example trial sequence, as well as examples of possible oddball targets which could replace one of the eight items in a sequence, shown in the bottom left corner of the panel. Panel A: an example visual sequence for the LowSTM Visual condition, and the single oddball target (an especially bright square). Panel B: an example visual sequence for the HighSTM Visual condition and the two oddball targets (an especially bright square or an especially dark square). Panel C: an example auditory sequence for the LowSTM Auditory condition and the single oddball target (an especially high pitch). Panel D: an example auditory sequence for the HighSTM Auditory condition and the two oddball targets (an especially high pitch or an especially low pitch). Panel E: an example audio-visual sequence for the LowSTM Audio-Visual condition and the two oddball targets (an especially high pitch or an especially bright square). Panel F: an example audio-visual sequence for the HighSTM Audio-Visual condition and the four oddball targets (an especially bright square, an especially dark square, an especially high pitch, or an especially low pitch).

Figure 5

Topographies of log-transformed alpha (top) and theta (bottom) power for each of the six conditions (left: Auditory, middle: Visual, right: Audio-Visual; alternating rows for LowSTM (one target) and HighSTM (two targets) variations) during the first second of the stimulus. Black circles mark electrode clusters of interest chosen a priori.

Figure 6

Time-frequency transforms for each of the six conditions (Columns: Auditory, Visual and Audio-Visual, Rows: LowSTM (one target) and HighSTM (two targets)) in Experiment 2, averaged over the same posterior cluster of electrodes as in Experiment 1. Dotted white line represents stimulus onset; solid white line represents one second after stimulus onset. Horizontal black lines bracket the alpha frequency band (8–13 Hz).

Figure 7

Time-frequency transforms for each of the six conditions (Columns: Auditory, Visual and Audio-Visual, Alternating Rows: LowSTM (one target) and HighSTM (two targets), Top Half: Power scale from 4.5–6.5 to highlight the difference in theta power among conditions) in Experiment 2, averaged over the same anterior cluster of electrodes as in Experiment 1. Bottom Half: Power scale from 4.2–6.8, used in all other plots of oscillatory power. Dotted white line represents stimulus onset; solid white line represents one second after stimulus onset. Horizontal black lines represent the theta frequency band analyzed (4–7 Hz).

Figure 8

Time-frequency transforms of induced oscillations after a subtraction of the averaged waveform for each subject at each electrode from each trial in Experiment 2. Columns: Auditory, Visual and Audio-Visual, Alternating Rows: LowSTM (one target) and HighSTM (two targets), Top Half: Averaged activity over the posterior cluster of electrodes, Bottom Half: Averaged activity over the anterior cluster of electrodes. Dotted white line represents stimulus onset; solid white line represents one second after stimulus onset.

Figure 9

Inter-trial phase coherence within the posterior (top half) and anterior (bottom half) clusters of electrodes for each condition (first and third rows: LowSTM Auditory, LowSTM Visual, LowSTM Audio-Visual; second and fourth rows: HighSTM Auditory, HighSTM Visual, HighSTM Audio-Visual). Dotted white lines represent stimulus onset; solid white lines represent one second after stimulus onset.

Figure 10

Left Column: ERPs for the anterior cluster of electrodes for the Auditory (top panel), Visual (middle panel) and Audio-Visual (bottom panel) conditions, for the LowSTM (blue) and HighSTM (red) conditions. Blocked conditions prevented common baselining for all three conditions. The vertical bar at 0 seconds represents stimulus onset. Right: Periodograms depict the mean amplitude of oscillations at particular frequencies (1–20 Hz) for the Auditory (top panel), Visual (middle panel) and Audio-Visual (bottom panel) conditions, for both the LowSTM (blue) and HighSTM (red) conditions.