Busse et al. 10.1073/pnas.0507704102. |
Fig. 5. Control ERP experiment for visual–visual interactions. (A) Stimuli and task. Streams of brief, unilateral, visual stimuli (Gabor gratings) were rapidly presented to the lower left and lower right visual quadrants while subjects attended to one lateral stream or the other. The location of covert attention is depicted here by the red dashed circles. Half of these lateral visual stimuli were accompanied by a task-irrelevant midline grating stimulus (below fixation), analogous to the task-irrelevant midline auditory stimulus in the multisensory audiovisual paradigm described in the article. Thus, in this visual–visual paradigm, sometimes this midline visual stimulus was presented simultaneously with either an attended lateral stimulus (Upper) or an unattended lateral stimulus (Lower). (B) Extracted ERPs for the irrelevant midline stimuli, displayed at three midline electrode locations. Pink traces depict extracted ERP responses for midline stimuli paired with an attended lateral visual stimulus, and black traces depict the extracted ERPs associated with the irrelevant midline stimuli paired with an unattended lateral visual stimulus. As can be seen, there was little difference between these extracted ERP responses for the irrelevant midline visual stimuli as a function of being synchronous with an attended versus an unattended lateral visual stimulus and, in particular, no suggestion of the long-latency sustained activity enhancement observed for the task-irrelevant midline auditory stimulus in the main experiment.
Supporting Text
In the experiment described in the main manuscript, it was observed that a task-irrelevant midline tone received enhanced processing when it occurred synchronously with an attended versus an unattended lateral visual stimulus. From these results, it was inferred that attention to one sensory modality can spread to encompass simultaneous signals from another modality, even when they are task-irrelevant and from a different location. It was suggested that this cross-modal attentional spread reflects an object-based, late selection process wherein spatially discrepant auditory stimulation is grouped with synchronous attended visual input into a multisensory object, resulting in the auditory information being pulled into the attentional spotlight and given enhanced processing.
As mentioned in the article, it could be argued that the enhanced late activity for the irrelevant midline tone was not specific for the auditory stimulus and does not necessarily reflect its grouping with the synchronous attended visual stimulus. In particular, it is conceivable that the occurrence of an attended visual stimulus might lead to a brief period of increased nonspecific neural responsiveness, such that the processing of any stimulus that occurs simultaneously with the attended stimulus would tend to be enhanced. To rule out this possibility, we performed a control event-related potentials (ERP) study examining for analogous visual–visual interactions. As in the multisensory study, subjects attended to either a left or a right stream of unilateral visual stimuli, but in this case, half of the lateral visual stimuli were accompanied by an irrelevant visual stimulus in the midline rather than an irrelevant auditory stimulus.
Methods
Subjects.
Sixteen members of the Duke University community (10 male and 6 female, ranging from 20 to 35 years of age) participated in the study. The study protocol was approved by the Duke University Health System Institutional Review Board, and informed consent was obtained from each participant.Stimuli and Task.
Subjects were seated in a sound-attenuated and dimly lit recording chamber facing a computer monitor on which all stimuli were presented. Stimulus display and response collection were performed by using the presentation software package (Neurobehavioral Systems, Albany, CA) running on a personal computer. In different blocks (»2.25 min in duration), subjects fixated a central cross while covertly directing their visual attention to either the left or right side of the monitor. Streams of brief (34 ms), unilateral, visual stimuli were rapidly presented [stimulus onset asynchronies (SOAs) = 350-550 ms] to the lower left or lower right visual quadrants (1° below the fixation point, 2° to the left or right) (Fig. 5A). The visual stimuli were Gabor patches, vertically oriented six-cycles-per-degree cosine gratings with a two-dimensional Gaussian envelope with a standard deviation of 0.15°. Subjects pressed a button with their right index finger upon detecting infrequent (»10% probability) target stimuli occurring in the stream at the attended location. Targets were gratings tilted slightly off vertical. Half of the lateral stimuli were accompanied by an additional grating stimulus presented in the midline directly below (2.24° below) the fixation point.Subjects were instructed to attend to the visual stimuli on the designated side on each run to detect the occasional target stimulus in that stream and to ignore all of the stimuli (both standards and targets) outside of the attended location. Reaction times (RTs) and accuracy of the responses to the targets were recorded for all of the runs. The target difficulty was titrated for each subject so that detection required highly focused attention but could be performed at a »80% correct performance criterion. This adjustment was accompanied by slightly changing the difference angle between the standard (vertical) and target (slightly oblique) gratings.
ERP Recording and Analysis.
ERP recording methods were identical to those described in the article. Using analogous subtractions to those in the multisensory experiment described in the article, the ERP response to the task-irrelevant midline visual stimulus was extracted as a function of whether it occurred simultaneously with an attended versus an unattended lateral visual stimulus. More specifically, extracted ERP responses for midline stimuli paired with an attended lateral visual stimulus were derived by subtracting the ERPs to the combined occurrence of an attended lateral stimulus and an irrelevant midline stimulus minus the ERPs to the attended lateral stimulus presented alone. Analogously, extracted ERPs associated with the irrelevant midline stimuli paired with unattended lateral visual stimulus were derived by subtracting ERPs to the combined occurrence of an unattended lateral stimulus and an irrelevant midline stimulus minus the ERPs to the unattended lateral stimulus presented alone. ANOVAs were performed contrasting these attentional context difference waves, analogous to the analyses performed on the extracted tone responses in the article.Results
As expected, prominent effects of visuo-spatial attention on the ERP responses to the lateral visual stimuli were observed, with larger P1 and N1 waves over occipital and parietal-occipital scalp sites when attended (P values ranging from .008 to .0001). In addition, targets in the attended lateral visual stream, but not in the unattended one, elicited large P300 waves associated with target detection. These results indicate the effectiveness of the spatial attentional manipulation.
On half the trials, an additional, task-irrelevant midline grating stimulus could appear simultaneously with a grating stimulus in an attended lateral location or in an unattended lateral location. To examine the effect on these synchronous midline visual stimuli of attending to the lateral visual stimuli, ERP responses evoked by single stimuli presented in either the left or the right location were subtracted from the responses to the corresponding stimulus pair (e.g., a left and midline stimulus minus a left stimulus alone), separately for each attention condition. In this way, the ERPs associated with the irrelevant midline visual stimulus as a function of its attentional context could be extracted.
Fig. 5B shows extracted ERPs at frontal, central, and occipital scalp sites for the irrelevant midline visual stimuli as a function of when they occurred with an attended or an unattended lateral visual stimulus. As can be seen, these responses showed little difference as a function of these multivisual-stimulus attentional contexts. Most directly analogous to the multisensory experiment in the article, there was no long-latency sustained negativity (or positivity) either at the frontal sites or any other scalp sites, including the ones over the occipital (visual) cortex. An analysis of variance of the mean amplitudes in the latency range of the processing negativity observed in the main study (220–700 ms) confirmed this observation (F1,15 = 1.32, P = 0.27).
Discussion
If the enhancement of processing for the task-irrelevant midline auditory stimuli observed in the multisensory experiment in the article was the result of a nonspecific increase in neural responsiveness in conjunction with the occurrence of a simultaneous attended visual stimulus, the extracted ERPs for the task-irrelevant midline visual stimuli in this visual–visual control experiment should have displayed an enhancement similar to that seen for auditory stimuli in the multisensory experiment. However, unlike the robust, long-latency, sustained negativity seen in the extracted auditory ERP responses as a function of multisensory attentional context, the task-irrelevant midline visual stimuli in this control experiment showed no such differences as a function of occurring synchronously with an attended versus an unattended lateral visual stimulus. Thus, these results indicate that the enhancement of the auditory tone response observed in the main experiment as a function of its multisensory attentional context is not the result of a transient increase in nonspecific neural responsiveness. Rather, these results indicates that this effect is specific to the stimulus configuration being multisensory in nature, and that it is a reflection of the specific grouping of the auditory stimulus with the simultaneously attended visual stimulus and the spread of attention to that auditory stimulus.