Roles of Speed vs. Inhibition-Concentration in Speech Recognition of Mandarin Cochlear Implant Users

Neurocognitive factors contribute to heterogeneity in cochlear implant (CI) success. This study addresses three gaps: (1) whether long, low-redundancy sentences in quiet vs. short, high-redundancy sentences in noise recruit different cognitive involvement; (2) the methodological conflation of processing speed with inhibition-concentration; (3) the scarcity of tone language evidence in understanding the cognitive-speech relationship.

Twenty-five postlingually deafened Mandarin CI users completed long-sentence recognition in quiet (Mandarin Chinese Adaptation of AzBio Sentence, CMnBio) and short-sentence recognition in four-talker babble noise (the BKB Sentences in Noise, BKB-SIN). Auditory input quality was assessed using the Spectral-Temporally Modulated Ripple Test (SMRT). The Shape Trail Test, a pencil-and-line task, was used to assess processing speed (STT-A/B), and inhibition-concentration (STT-B). Because STT-B reflects both processing speed and inhibition-concentration, the latter was isolated using STT-B.A (the residual of STT-B on STT-A). Independent contributions to recognition were assessed with hierarchical regression analyses.

SMRT, STT-A, and STT-B showed correlations with both sentence outcomes, whereas STT-B.A correlated solely with BKB-SIN. In regression models controlling for SMRT, STT-A explained significant variance in CMnBio (β = -0.507, p = 0.001) and BKB-SIN (β = -0.371, p = 0.012). STT-B.A contributed only to BKB-SIN (β = -0.318, p = 0.014).

Information processing speed shows a general association with sentence recognition in both quiet and noise, whereas inhibition-concentration shows a noise-specific association in Mandarin CI users. This study highlights the importance of methodological distinction separating these two cognitive functions and extends cognition-speech links to tonal languages.