Introduction

Conventional fMRI uses gradient-recalled echo echo-planar imaging (GRE-EPI), which often creates acoustic noise up to 120dBA¹, presenting challenges in auditory task fMRI by potentially confounding results. Attempts to reduce these effects haven't been fully successful, and GRE-EPI is prone to various artifacts affecting the spatial accuracy of neural activity localization². "Looping Star," a new silent fMRI technique, shows promise with lower noise and geometric distortion^{3, 4}. This study evaluates Looping Star's efficacy in auditory fMRI against standard methods.

Methods

The study protocol was approved (SH9H-2021-T449-1) by the Ethics Committee of Shanghai Jiao Tong University School of Medicine Affiliated Ninth People’s Hospital (Shanghai, China), and written informed consent was obtained from all participants. Seven healthy participants (three females; mean age=33.29 years ± 7.74 years; all right-handed) were recruited.
The experiment utilized a block-design passive listening task (Figure 1). Participants experienced six task blocks, each lasting 20 seconds, interspersed with seven rest blocks of 15 seconds each. During task blocks, subjects were exposed to a continuous stream of multilingual babble sourced from standard corpora commonly utilized in audiological assessments. This paradigm was consistently employed for both Looping Star and GRE-EPI image acquisitions. The order of Looping Star and GRE-EPI scans was counterbalanced among participants.
Imaging data were acquired on a 3.0-T scanner (Premier, GE Healthcare, WI) using 48-channel receive-only head coil. The parameters for the two-echo Looping Star and conventional GRE-EPI sequences were closely matched. For Looping Star: TE = 0 ms, 26.88 ms, TR = 2611 s, number of volumes = 90, matrix size = 64x64x64, resolution = 3mm, FOV = 19.2 cm, 24 spokes per echo, 2 echoes, flip angle=3°. For GRE-EPI: TE = 30 ms, TR = 2611 ms, 90 volumes, matrix size = 64x64, slice thickness = 3mm, 48 slices, slice gap = 0 mm, FOV = 19.2 cm, flip angle=90°. Additionally, a 3D T1-weighted Bravo image was acquired (TE =2.2 ms, TR =5.8 ms, TI = 450 ms, flip angle=12°, matrix size = 256x256, resolution = 1mm, 160 slices).
The preprocessing of Looping Star and GRE-EPI data followed a similar, yet slightly different pipeline. For GRE-EPI: the initial five volumes were discarded to address signal instability, followed by head motion correction using FSL's MCFLIRT, slice timing correction with AFNI's 3dTshift, a two-step registration process with ANTs, and finally, smoothing using an 8 mm full width at half maximum (FWHM) kernel. In contrast, Looping Star preprocessing involved both Echo 1 and Echo 2, with Echo 2 showing the BOLD effect and being utilized for further activation analysis. Unlike the GRE-EPI protocol, Looping Star did not require slice timing correction. In the registration phase, average images from Echo 1 were aligned to individual T1-weighted images and subsequently to standard space. These transformations were then applied to Echo 2. All other preprocessing steps were consistent with those used for GRE-EPI.
Individual auditory responses and group-level statistics were computed with SPM12. Multiple comparison was performed for each sequence (voxel p < 0.001, cluster p<0.05, cluster-level family-wise error corrected).

Results

Looping Star showed broader activation and stronger signal detection compared to GRE-EPI, with a notable voxel count increase in both hemispheres (Table 1, Figure 2). Looping Star demonstrates enhanced reliability over GRE-EPI in detecting activations within Heschl's gyrus (HG) and the superior temporal gyrus (STG) on both sides, which serve as the primary and associative auditory centers essential for auditory processing. Thus, Looping Star demonstrates superior sensitivity in neural activity detection over GRE-EPI, possibly due to its enhanced reduction of acoustic noise interference, yielding clearer activation patterns.
In addition, Looping Star uncovers activations in brain regions not detected by GRE-EPI, specifically in the left posterior central gyrus, right Rolandic operculum, and right insula (Table 1). Although the posterior central gyrus is primarily associated with somatosensory processing, its activation may reflect its integration with auditory functions⁵. The Rolandic operculum and the insula have been implicated in a variety of functions including language processing, emotional responses to sounds, and complex auditory processing that goes beyond simple sound perception^6,7. The silent acquisition characteristic of Looping Star may thus provide a more comprehensive view of the auditory processing network.

Conclusions

This preliminary study suggests that Looping Star improves the detection of auditory-related brain regions during listening task fMRI, owing to its reduction of acoustic interference. This technique holds promise for offering more detailed and extensive insights into auditory processing.

Acknowledgements

This work is supported by Science and Technology Commission of Shanghai Municipality Major Basic Research (2018SHZDZX05), Science and Technology Commission of Shanghai Municipality Shanghai Key Laboratory of Translational Medicine on Ear and Nose diseases (14DZ2260300), Shanghai Municipal Health Commission Shanghai Key Clinical Specialty Construction - Otolaryngology-Head and Neck Surgery (shslczdzk00802), Shanghai Shen Kang Hospital Development Center Emerging Frontier Project (SHDC12020105), Shanghai Jiao Tong University School of Medicine Translational Medicine Collaborative Innovation Project (TM202011), and Shanghai Jiao Tong University School of Medicine High-level Local University Construction Project.