Evoked and intrinsic asymmetries during auditory attention: implications for the contralateral and neglect models of functioning

Abstract

Unlike the visual system, a direct mapping of extrapersonal space does not exist within human auditory cortex (AC). Thus, models (contralateral bias vs. neglect) of how auditory spatial attention is allocated remain debated, as does the role of hemispheric asymmetries. To further examine these questions, 27 participants completed an exogenous auditory orienting task while undergoing functional magnetic resonance imaging. Resting-state data were also collected to characterize intrinsic activity within the AC. Current results provide the first evidence of hemispheric specialization in the "where" (right secondary AC) auditory processing stream during both evoked (orienting task) and intrinsic (resting-state data) activity, suggesting that spontaneous and evoked activity may be synchronized by similar cortical hierarchies. Strong evidence for a contralateral bias model was observed during rapid deployment stages (facilitation) of auditory attention in bilateral AC. However, contralateral bias increased for left and decreased for right AC (neglect model) after longer stimulus onset asynchronies (inhibition of return), suggesting a role for higher-order cortical structures in modulating AC functioning. Prime candidates for attentional modulation include the frontoparietal network, which demonstrated right hemisphere lateralization across multiple attentional states.

Figure 1.

This figure presents predictions and cartoon illustrations of the contralateral (Panel A) and neglect (Panel B) models of response to lateralized auditory stimuli. In the contralateral model, both left and right hemispheres display a bias (arrows) toward stimuli in the contralateral hemifield. Bar graphs indicate hypothesized greater activation (PSC, percent signal change) to contralateral stimuli (LS, left stimulus; RS, right stimulus) within each AC (L AC, left auditory cortex; R AC, right auditory cortex). In the neglect model, the L AC shows a bias toward stimuli in the right hemifield (gray area), while the R AC responds more equally to stimuli in both hemifields (white area). Panel C presents the trial structure for sample valid (Val-R) and invalid right (Inv-R) trials. For valid trials, cue (C) and target (T) occur in same headphone/hemifield; for invalid trials, targets occur in opposite headphones/hemifields.

Figure 2.

Panel A displays regions demonstrating Laterality × Validity interaction effects for 200 and 700 ms SOA in the left and right figures, respectively. Clusters of significant activation were present in the AC bilaterally at 200 ms SOA and in the left AC only at 700 ms SOA. Color scale indicates voxel-level significance (black, P < 0.005; white, P < 0.001). Panel B presents mean PSC for valid left (Val-L, white), valid right (Val-R, black), invalid left (Inv-L, light gray), and invalid right (Inv-R, dark gray) conditions for clusters within the AC. Panel C displays a sample of automatic parcellation (black areas) of the transverse temporal gyrus (TTG) in one subject, whereas Panel D presents mean group PSC within the TTG for the 4 trial conditions. Slices are located at 9 mm superior to the origin in Talairach space. All error bars represent 2 × standard error of the mean. Bracketed asterisks indicate significant findings (P < 0.05).

Figure 3.

Auditory regions (Panel A) that demonstrated a Hemisphere × Laterality interaction effect at 200 and 700 ms SOA. Color scale indicates voxel-level significance (black, P < 0.005; white, P < 0.001). Panel B presents mean PSC for valid left (Val-L) or valid right (Val-R) trials in left hemisphere (LH; white or black, respectively) and valid left or valid right in right hemisphere (RH; light gray and dark gray, respectively). Slices are located at 9 mm superior to the origin in Talairach space. All error bars represent 2 × standard error of the mean. Bracketed asterisks indicate significant findings (P < 0.05).

Figure 4.

Hemispheric asymmetries in functional activation. Greater right (RH) compared with left (LH) hemisphere activation (cool colors) was observed in dorsolateral prefrontal cortex (DLPFC), superior and inferior parietal area (SPL/IPL), PT and superior temporal gyrus (STG), pre-SMA, precuneus and posterior cingulate cortex (Precun/PCC), and cerebellar areas. Greater left compared with right hemisphere activation (warm colors) was observed in the sensorimotor areas, primary auditory cortex (PAC), and SMA/pre-SMA. Color scale indicates the number of analyses with overlapping areas of significant activation (dark blue and red, 1–3 trial conditions; cyan and yellow, all 4 trial conditions) Slices are located at x = −42 or −3 mm to the left of the origin in Talairach space.

Figure 5.

This figure presents regions that exhibited differences in intrinsic activity based on seed placement (Panel A) in the primary (PAC, purple seed) versus secondary (SAC, green seed) AC. Panel B presents hemisphere × seed laterality interaction effects for PAC (left column) and SAC (right column) seed. Bar graphs represent mean z-scores (error bars = 2 × standard error of the mean) for each of the 4 conditions (red, S_LH_L; blue, S_RH_R; orange, S_LH_R; green, S_RH_L) within clusters in the PAC, SAC and insula (Ins). Panel C presents main effects of seed laterality for PAC seed analyses (left column) and SAC seed analyses (right column). Regions that exhibited higher connectivity with the right seed (S_R) are presented in cool colors, whereas regions that exhibited higher connectivity with the left seed (S_L) are presented in warm colors. Color scale indicates voxel-level significance (dark blue and red, P < 0.005; cyan and yellow, P ≤ 0.001). Slices are located at x = −39 (PAC seed), x = −48 (SAC seed and Panel C) or x = −41 mm (Panel B) to the left of the origin in Talairach space.