Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Grounding inner speech diversity in the dynamics of speech production-perception control
Poster Session D, Thursday, October 1, 4:30 - 6:30 pm, Wangari Maathai
Helene Loevenbruck1, Monica Baciu1,2; 1Univ. Grenoble Alpes, Univ. Savoie Mont Blanc, CNRS, LPNC, 38000, Grenoble, France, 2Neurology Department CMRR, Grenoble Alpes Hospital, Grenoble, France
Inner speech, or endophasia, is the silent mental production of language, often conceptualized as overt speech without articulation. Behavioural and neuroimaging evidence indicates that inner speech recruits auditory, articulatory, somatosensory, and visual systems similarly to overt speech [1,2]. Auditory involvement is supported by verbal transformation effects during covert repetition, perceptual biases induced by internally generated syllables, and activation of auditory cortices. Motor and somatosensory contributions are evidenced by increased lip muscle activity, respiratory asymmetries, and activation in premotor, motor, and somatosensory cortices. Visual cortex activation has also been reported during dialogal inner speech and sign maintenance in short-term memory. Despite these shared mechanisms, endophasia differs from overt speech in important respects. A first defining feature is condensation: syntax may be abbreviated, semantics compressed, and phonology reduced, yielding experiences lacking articulatory or auditory qualities [3]. Behaviourally, inner speech is often faster and more fluent than overt speech, as shown by fewer errors during covert tongue-twister tasks [2]. The tip-of-the-tongue phenomenon and auditory verbal aphantasia further suggest that endophasia can occur without full phonetic specification [4,5], supporting a continuum from fully expanded, multisensory forms to fully condensed, amodal experiences. A second distinguishing dimension is dialogality. Unlike overt speech, typically experienced in one’s own voice, endophasia may be monologal or dialogal, involving internally simulated other voices. Neuroimaging studies show that dialogal inner speech recruits additional right-lateralized frontal and parietal regions, as well as the precuneus and posterior cingulate cortex, likely reflecting voice monitoring and perspective switching [6,7]. A third dimension is intentionality: whereas overt speech is generally intentional and controlled, endophasia may be intentional (e.g., verbal rehearsal) or unintentional (e.g., mind-wandering), sometimes with reduced control [7,8]. Furthermore, dissociations observed in stuttering [9], non-fluent aphasia [10], and L2 learners [11], where endophasia may remain fluent despite overt speech difficulties, have been interpreted as evidence for distinct neural mechanisms. The ConDialInt model [7] resolves this paradox by proposing a shared neurocognitive architecture for inner and overt speech. Here, we present its revised version which further specifies the role of perceptual subprocesses in inner speech awareness and clarifies the hierarchical production-perception dynamics. Drawing on predictive control accounts, the model posits that inner speech emerges from the same production–perception loop as overt speech. During overt speech, predictive mechanisms operate across multiple levels, from conceptualization to motor execution, by generating expected outcomes from copies of intermediate production signals and evaluating them through corresponding perceptual subprocesses before articulation. Inner speech arises when this process is interrupted before execution, rendering generated predictions consciously accessible. This explains condensation (late vs. early interruptions preserve or bypass sensory predictions, yielding expanded or condensed forms), dialogality (the predictive system, tuned for both production and perception, generates self- and other-voices), and intentionality (higher-level control over initiation and interruption). Recent sEEG evidence showing that motor planning precedes sensory prediction in both overt and inner speech, supports this account [12]. Thus, endophasia is not a distinct language system, but an exaptation of speech production and perception control mechanisms.
Topic Areas: Speech Motor Control, Language Production