اثرات عصبی حواس پرتی شنوایی بر توجه بصری
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|39114||2013||8 صفحه PDF||سفارش دهید||6283 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Brain and Cognition, Volume 81, Issue 2, March 2013, Pages 263–270
Abstract Selective attention in the presence of distraction is a key aspect of healthy cognition. The underlying neurobiological processes, have not, however, been functionally well characterized. In the present study, we used functional magnetic resonance imaging to determine how ecologically relevant distracting noise affects cortical activity in 27 healthy adults during two versions of the visual Sustained Attention To Response Task (SART) that differ in difficulty (and thus attentional load). A significant condition (noise or silence) by task (easy or difficult) interaction was observed in several areas, including dorsolateral prefrontal cortex (DLPFC), fusiform gyrus (FG), posterior cingulate (PCC), and pre-supplementary motor area (PreSMA). Post hoc analyses of interaction effects revealed deactivation of DLPFC, PCC, and PreSMA during distracting noise under conditions of low attentional load, and activation of FG and PCC during distracting noise under conditions of high attentional load. These results suggest that distracting noise may help alert subjects to task goals and reduce demands on cortical resources during tasks of low difficulty and attentional load. Under conditions of higher load, however, additional cognitive resources may be required in the presence of noise.
Introduction The ability to selectively attend to relevant stimuli in the environment while ignoring salient distractors is a key component of cognition. Thousands of drivers are killed each year in the United States due to “inattention” (www.census.gov), including accidents caused by environmental distraction. Attention deficit hyperactivity disorder, for which symptoms include being easily distracted (Wolraich, 2006), exerts a “significant negative impact of quality of daily life, including work, social life, and relationships” (Rösler, Casas, Konofal, & Buitelaar, 2010). In schizophrenia, the inability “to choose which sources of input should be attended” is a debilitating cognitive symptom and has recently been suggested as an essential biomarker for treatment intervention (Luck, Ford, Sarter, & Lustig, 2012). To a large extent, the effect of irrelevant noise on task performance has been investigated in terms of its deleterious consequences. Early work by Broadbent found that noise interfered with vigilance (Broadbent, 1951), response speed (Broadbent, 1957), and mental arithmetic (Broadbent, 1958). Similar findings have since been reported using many different tasks and subject populations (Smith, 1989). However, irrelevant noise may also facilitate performance, particularly during monotonous, repetitive tasks (Smith, 1989 and Suter, 1989). Indeed, as initially proposed by Yerkes and Dodson (1908) and expanded upon by Zentall and Zentall (1983), the effect of task-irrelevant stimuli may depend on task difficulty, as appropriate levels of stimulation may be required for optimum performance. The neuronal processes that underlie the proposed interaction between task difficulty and distraction have not been well characterized. Insight into this process can be gained, however, by examining two variants of an attention task, the Sustained Attention to Response Task (SART), which may be comparable except for their respective difficulties. In the SART, a subject is shown a series of numbers and instructed to press a button when he sees any number except for “3”; if he sees a “3,” the subject is instructed to withhold from responding. The task has a much lower frequency of “3’s” than other numbers; thus the primary task objective is to inhibit the natural tendency to button press after each stimulus presentation (O’Connell et al., 2004). The Fixed version of the SART, in which targets are presented (one at a time) in numerical order and the stop target is therefore predictable, is easy and requires fewer trial-by-trial attentional resources. The Random version, in which numbers are presented in random order, requires the subject to more fully process each stimulus to perform the task accurately and is therefore more difficult, requiring higher attentional load. Behavioral evidence for increased processing is an increase in reaction time for the Random SART. fMRI studies have characterized the functional neuroanatomy of the Fixed and Random SART. Both versions have shown involvement of a frontal–parietal attention network that includes the dorsolateral prefrontal cortex (DLPFC) and inferior parietal lobule (Fassbender et al., 2004). DLPFC activation in particular is associated with top-down control processes as demonstrated via neuroimaging in other attention tasks (Banich et al., 2000, Cohen et al., 1998, Hager et al., 1998 and Sturm et al., 1999). The Random SART showed additional activity in the inferior frontal gyrus and basal ganglia, likely reflecting its additional demands on response inhibition. The Random SART also showed increased activity in visual cortex, reflecting its demands on sensory processing (Fassbender et al., 2004). A later imaging study used randomly presented “alerting tones” during the Random SART, and showed deactivation of the frontal portion of the attention network despite no change in performance (O’Connell et al., 2011). The decreased activation in this area was interpreted to reflect decreased need for top-down attentional control due to the cueing, alerting effect of the tones. In addition, increased activity in the left DLPFC with tones was observed during a control task in which the subject was instructed to press after every stimulus, suggesting an “orienting” response to the tones ( Corbetta and Shulman, 2002 and O’Connell et al., 2011). Decreased activity in the DLPFC with tones during the Random SART suggests that exogenous stimulation may reduce demands on cortical attention networks. However, a number of questions remain. The tones in the O’Connell et al. (2011) study were task-relevant; it is unclear if task-irrelevant stimulation would have the same effect on Random SART-associated activity. In addition, the neurobiological effects of task-irrelevant auditory stimulation during tasks of comparatively low difficulty, such as the Fixed SART, are unknown. Finally, the alerting tones in the O’Connell et al. (2011) study were intermittent (every 8–12 s) and of the same frequency (2 kHz) and duration (30 ms); it is unclear whether constant noise would have the same effect, or conversely increase the burden on attentional processing. Indeed, based on previous studies that examine the effect of cross-modal distraction on attention, one might predict that constant noise would increase response in areas crucial for processing the attended modality (Langner et al., 2011 and Roland, 1982). In the present study, using functional magnetic resonance imaging (fMRI), we compared the neurophysiological effects of task-irrelevant “urban noise” stimulation on the Random and Fixed SART. The “urban noise” is a mixture of talk radio, music, and conversation one might find on a crowded city street, and is designed to mimic real-world sounds (Tregellas, Ellis, Shatti, Du, & Rojas, 2009). We formulated two hypotheses: (1) Relative to Fixed SART, Random SART would additionally recruit areas important for attentional, inhibitory, and sensory processing, because the Random version is more difficult and requires more resources than the Fixed version; (2) a significant Task × Noise interaction would be observed in areas important for attentional and sensory processing (e.g. the DLPFC and visual cortex), suggesting that the effect of noise may differ depending on attentional load.
نتیجه گیری انگلیسی
. Results 3.1. Effect of noise on fixed and random SART performance As anticipated, no significant difference was observed between errors of commission on the Fixed SART in silence and noise (t = 0.63, df = 26, p = 0.53) or the Random SART in silence and noise (t = 0.40, df = 26, p = 0.69). No significant difference was observed between omission errors in the Fixed SART in silence and noise (t = 0.12, df = 26, p = 0.90) or the Random SART in silence and noise (t = 0.18, df = 26, p = 0.85). Likewise, no significant difference was observed between reaction times in the Fixed SART in silence and noise (t = 0.31, df = 26, p = 0.76) or the Random SART in silence and noise (t = 0.11, df = 26, p = 0.91) ( Table 1a and Table 1b). Table 1a. Behavioral data, Fixed SART. Measure Fixed silent Fixed noise p % Errors of commission 4.42 ± 1.01 3.69 ± 1.04 0.50 % Errors of omission 4.13 ± 1.63 3.28 ± 1.37 0.31 Reaction time (ms) 297 ± 7.87 289 ± 8.38 0.13 Table options Table 1b. Behavioral data, Random SART. Measure Random silent Random noise p % Errors of commission 22.55 ± 3.71 21.52 ± 3.77 0.78 % Errors of omission 1.03 ± 0.084 0.92 ± 0.52 0.74 Reaction time (ms) 374 ± 15.36 373 ± 14.41 0.44 Table options 3.2. fMRI results: Random > Fixed SART, collapsed across all conditions To test the hypothesis that the Random SART requires more visual processing than the Fixed SART, as well as to determine which additional resources were required for the Random version, the contrast Random > Fixed SART (collapsed across noise and silence conditions) was examined. Significant activation was observed in primary and accessory visual cortex, cingulate gyrus, inferior frontal gyrus, insula, cerebellum, and thalamus (Fig. 2 and Table 2). Regions recruited for the contrast Random>Fixed SART, collapsed across silence ... Fig. 2. Regions recruited for the contrast Random > Fixed SART, collapsed across silence and noise conditions. Statistical parametric maps thresholded at p < 0.001 for visualization. Images displayed in neurologic convention (R on R). Figure options Table 2. MNI coordinates and statistics for the contrast Random > Fixed SART, collapsed across all noise and silence conditions. Regions grouped by contiguous clusters. T values drawn from local maxima. Brain Areas Hemi Brodmann area t value x, y, z (mm) Cluster (voxel size) Fusiform gyrus L 19 9.16 −39, −64, −8 2395 Middle occipital gyrus L 18 9.07 −39, −91, 1 Middle occipital gyrus L 18 8.47 −30, −97, −2 Middle temporal gyrus R 37 7.83 45, −64, 11 1594 Superior temporal gyrus R 22 7.69 39, −58, 11 Lingual gyrus R 17 7.28 21, −91, 4 Inferior frontal gyrus R 45 7.37 42, 23, 4 2119 Inferior fronal gyrus R 47 7.13 33, 26, −2 Inferior frontal gyrus R 47 7.03 36, 20, −8 Inferior frontal gyrus L 45 6.70 −30, 26, 1 469 Basal ganglia L 4.61 −18, 2, −11 Anterior cingulate R 32 5.80 9, 23, 31 716 Thalamus R 5.65 3, −31, 1 371 Thalamus L 5.59 3, −13, 10 Table options 3.3. fMRI results: Fixed > Random SART, collapsed across all conditions Whole brain analysis of the contrast Fixed > Random SART revealed significant activation of the precuneus (peak (x, y, z) 9, −70, 52, t = 5.93, k = 243 voxels). 3.4. fMRI results: Interaction contrast To test the hypothesis that task-irrelevant noise stimuli is associated with contrasting effects on Random and Fixed SART, the interaction contrast (noise x task) was examined. Whole-brain analysis revealed significant interaction effects in middle frontal gyrus, medial frontal gyrus/pre-supplementary motor area (PreSMA), cerebellum, fusiform gyrus, midbrain and posterior cingulate (Fig. 3A–C and Table 3). Brain regions showing significant response for the interaction contrast (Random ... Fig. 3. Brain regions showing significant response for the interaction contrast (Random SART during noise > Random SART during silence > (Fixed SART during noise > Fixed SART during silence)). Statistical parametric maps thresholded at p < 0.001 for visualization. Images displayed in neurologic convention (R on R). Figure options Table 3. MNI coordinates and statistics for brain regions showing greater responses for the whole brain interaction contrast (Random SART during noise > Random SART during silence > (Fixed SART during noise > Fixed SART during silence)). Regions grouped by contiguous clusters. T values drawn from local maxima. Brain areas Hemi Brodmann area t Value x, y, z (mm) Cluster (voxel size) Cerebellum R 5.22 42, −64, −20 132 Fusiform gyrus R 19 4.76 33, −73, −14 Midbrain L 5.00 −3, −19, −8 113 Midbrain R 4.82 9, −13, −2 Basal ganglia R 4.35 21, −4, 1 Posterior cingulate R 29 4.97 6, −46, 16 119 Posterior cingulate L 29 3.91 −6, −49, 13 Pre-SMA L 6 4.37 −9, 5, 61 246 Pre-SMA R 6 4.04 6, −13, 64 Superior frontal gyrus L 8 3.78 −6, 17, 52 Table options To further explore these results, post hoc analyses of significant interaction effects were performed (Fig. 4). These tests revealed significantly decreased response in the fusiform gyrus during the FixedNoise condition relative to FixedSilent (t = 3.77, df = 26, p = 0.001) and a nearly significant increase in response in the fusiform gyrus during the RandomNoise condition relative to RandomSilent (t = 1.86, df = 26, p = 0.07). In the posterior cingulate, a nearly significant decrease in response was observed during the FixedNoise condition compared to FixedSilent (t = 1.94, df = 26, p = 0.06) and significantly increased response was observed during the RandomNoise condition relative to RandomSilent (t = 2.98, df = 26, p < 0.01). In the pre-SMA, significantly decreased response was observed in the FixedSilent condition relative to FixedNoise (t = 2.86, df = 26, p < 0.01) and no difference was observed between the RandomNoise and RandomSilent conditions (t = 1.17, df = 26, p = 0.25). Direction of interaction effects. In Fig. 4A–C, the Y-axis represents the ... Fig. 4. Direction of interaction effects. In Fig. 4A–C, the Y-axis represents the percent BOLD signal change (relative to baseline) extracted from an 8 mm sphere centered at the peak of significant clusters reported in Fig. 3. In Fig. 4D, the Y-axis represents the mean voxelwise percent BOLD signal change (relative to baseline) for the DLPFC anatomical ROI. Error bars represent the standard error. *p < 0.05; ^p < 0.1. Figure options 3.5. fMRI results: DLPFC ROI To test the hypothesis that task-irrelevant noise stimuli is associated with contrasting effects on Random and Fixed SART in the DLPFC, the mean voxelwise percent signal change was examined for all four task conditions from an anatomically-defined ROI of the left DLPFC. Using this analysis, a significant interaction effect was observed in the left DLPFC (t = 3.37, df = 26, p = 0.002; peak coordinate (x, y, z) −27, 38, 46; Fig. 3D). Post hoc analysis of the direction of this effect revealed a nearly significant decrease in response in the DLPFC during the FixedNoise relative to FixedSilent condition (t = 2.00, df = 26, p = 0.056) but no difference in response between the RandomNoise and RandomSilent conditions (t = 0.7, df = 26, p = 0.49).