Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Does rapid-speech exposure make standard-rate speech sound slower? Concurrent fNIRS and eye tracking during multimodal L2 sentence comprehension
Poster B69 in Poster Session B, Wednesday, September 30, 4:30 - 6:30 pm, Wangari Maathai and Lise Girardin
This poster is part of the Sandbox Series.
Mayumi Kajiura1; 1Nagoya City University
Speech comprehension requires listeners to adjust temporal expectations as speech rate changes. Previous work has shown rapid adaptation to time-compressed and naturally fast speech, including transfer across fast-speech conditions (Adank & Janse, 2009). Neuroimaging evidence further suggests that adaptation to time-compressed sentences involves auditory association and premotor systems (Adank & Devlin, 2010). From a temporal-prediction perspective, Kösem et al. (2018) showed that neural entrainment to preceding speech rhythms can persist after a rate change and bias the perception of subsequent words. Building on this account, the present study asks how short-term exposure to double-rate L2 speech changes the subjective and neurocognitive processing of subsequent standard-rate speech. This study examines whether short-term exposure to double-rate English speech leads subsequent standard-rate speech to be perceived as slower and easier to process during L2 sentence comprehension. During the rate-adaptation phase, speech and written sentences were presented concurrently to test whether learners rely on orthographic information as a compensatory resource when auditory comprehension is strained, especially under double-rate conditions. To clarify the mechanism underlying the subjective slowing effect, the study further tests whether speech-rate context and compensatory reliance on written input are reflected in response time, rereading/regression patterns, and prefrontal and bilateral temporal–parietal hemodynamic activity. Fifty-six participants completed an English sentence plausibility-judgment task while fNIRS and eye movements were recorded concurrently. Hemodynamic responses were recorded using an fNIRSport2 system with a 38-channel montage, including eight short-separation channels, covering prefrontal and bilateral temporal–parietal regions. Gaze behavior was recorded with a Tobii Pro Spectrum. Stimuli were arranged in two counterbalanced versions of a 140-trial protocol. The present analyses focus on the rate-adaptation phase, in which English speech and written sentences were presented concurrently at either standard or double speech rate. Each 10-trial block included 4–6 s jittered rest intervals, creating mini-blocks that allow comparison of the first and last three trials within a block. Subjective ratings of perceived speed, difficulty, and ease of comprehension were collected, along with digit span, Stroop performance, and English proficiency. Planned analyses will model accuracy, onset-locked response time, speech-offset-adjusted response latency, subjective ratings, rereading/regression-related gaze measures, and HbO/HbR responses as functions of current speech rate, preceding rate context, block position, mini-block position, and individual differences. Double-rate trials are expected to increase perceived speed and difficulty, rereading/regression patterns, and prefrontal HbO responses, reflecting greater control and repair demands when semantic integration cannot keep pace with the auditory stream. Critically, standard-rate trials following double-rate exposure are expected to feel slower and easier to process, with faster responses, fewer rereading/regression patterns, and reduced prefrontal recruitment. Bilateral temporal–parietal responses will be examined exploratorily as indices of auditory-language processing, auditory–orthographic integration, and speech-rate context effects. The expected pattern would support a predictive-processing account in which recent speech-rate context recalibrates temporal expectations, reducing visual repair and compensatory frontal engagement during subsequent standard-rate L2 comprehension. Data collection is complete; preprocessing and statistical modeling are ongoing.
Topic Areas: Multisensory or Sensorimotor Integration, Speech Perception