Listening Without the Noise: Near-Silent Looping Star fMRI Reveals Neural Processing of Degraded Speech

Excessive acoustic noise during fMRI poses challenges for auditory cognitive neuroscience. We aimed to evaluate a near-silent fMRI sequence, Looping Star (LS), for measuring brain responses to acoustically degraded speech. We hypothesized that LS would maintain BOLD sensitivity comparable to conventional Gradient-echo EPI and potentially reveal enhanced activation in brain regions associated with effortful speech processing under degraded listening conditions. Ten healthy adult native German speakers underwent fMRI while passively listening to blocks of spoken sentences under four conditions: clear speech, moderately degraded speech (spectral noise vocoding at ~50% intelligibility), unintelligibly degraded speech and clear speech mixed with EPI-like scanner acoustic noise. Each participant was scanned in four runs (two using standard single-echo EPI and two using the near-silent multi-echo LS sequence) in an alternating order. All fMRI timeseries (single-echo EPI, single-echo LS and echo-combined LS) detected robust bilateral auditory cortex activation for all Sound > silent baseline. Echo-combined LS, however, yielded larger activation clusters and higher peak effect sizes, extending beyond auditory cortex regions. Notably, only echo-combined LS revealed significant activation in the left inferior frontal gyrus (inferior frontal gyrus, pars opercularis) and adjacent regions when comparing all speech stimuli to baseline, whereas EPI and single-echo LS showed no such frontal activation. Direct between-sequence comparisons demonstrated significantly greater BOLD responses in echo-combined LS compared to EPI in key regions (including left insula, left precentral gyrus and left inferior frontal gyrus) for the all Sound > baseline and moderately degraded > baseline contrasts. These findings underscore the promise of near-silent fMRI for auditory neuroscience: by virtually eliminating scanner noise, LS fMRI can reveal neural responses to degraded speech that might be masked during loud EPI scans.