The DIVA Model of Speech Production & Its Implications in Stuttering

Fluent speech requires the brain to coordinate a rapid sequence of precisely timed motor actions. During normal conversation, speakers may produce several syllables per second, each involving coordinated movements of the tongue, lips, jaw, larynx, and respiratory system. These movements must occur in the correct order and with highly precise timing for speech to sound smooth and continuous. Even small disruptions in this coordination can alter the flow of speech and manifest itself in a variety of speech disorders.

Modern-day  neuroscience increasingly frames speech production as a sensorimotor control issue, where the brain continuously plans, initiates, monitors, and adjusts movements of the vocal tract while using sensory feedback to detect and correct errors  (Neef & Chang, 2024).  Within this framework, fluent speech emerges from the coordinated activity of multiple brain systems distributed across cortical and subcortical regions. Cortical language areas generate speech plans, motor cortex regions control articulatory movements, sensory systems monitor the resulting speech sounds, and subcortical circuits regulate the timing and initiation of motor commands. One of the most influential frameworks for understanding how these processes interact is the DIVA model (Directions Into Velocities of Articulators) developed by Tourville and Guenther, 2011. The DIVA model proposes that speech production relies on two interacting control systems: a feedforward control system and a feedback control system.

The feedforward system consists of learned motor commands that allow speakers to produce familiar speech sounds rapidly and automatically. Over years of speaking practice during childhood, the brain develops stable motor programs that encode how to produce specific sounds and syllables. When a speaker intends to produce a particular word, the feedforward system sends motor commands to articulatory muscles, allowing speech to unfold quickly without requiring constant conscious monitoring. Concurrently, a feedback control system monitors the sensory consequences of speech movements. Auditory and somatosensory systems compare the sounds and physical sensations produced during speaking with the brain’s predicted outcomes. If discrepancies arise, for example, if a sound is mispronounced, the feedback system can generate corrective signals that adjust subsequent movements in order to produce fluent and correct speech.

In fluent speakers, these two systems operate together efficiently. The feedforward system allows speech to occur rapidly and automatically, while the feedback system ensures accuracy by correcting occasional errors. Within this framework, subcortical circuits involving the basal ganglia and thalamus are thought to play an important role in initiating feedforward speech motor programs (Alm 2004).  These circuits form part of the cortico–basal ganglia–thalamocortical loop, which regulates when motor commands should be released (Chang & Guenther, 2020).

Because speech requires the repeated initiation of successive motor programs, the stability of these initiation signals is crucial for maintaining fluency. Each syllable in a sentence must be triggered at precisely the right moment in order for speech to flow smoothly from one segment to the next. Researchers have proposed that stuttering may involve disruptions in these initiation processes. In a neurocomputational account integrating the DIVA model with neuroimaging findings, Chang and Guenther (2020) suggested that developmental stuttering may reflect differences in how the cortico–basal ganglia–thalamocortical loop initiates speech motor programs. If the basal ganglia fail to generate stable timing cues for initiating these programs, the release of motor commands may become inconsistent. Under such conditions, the brain may attempt to initiate the same motor program multiple times, producing repetitions of sounds or syllables. In other cases, competing motor signals may prevent the articulatory movement from being released altogether, resulting in a temporary speech block in which the speaker attempts to speak but the movement fails to initiate.

The DIVA framework also helps explain why stuttering can vary across situations. When feedforward motor programs become less reliable, speakers may rely more heavily on feedback control systems to guide speech movements. Feedback-based corrections, however, occur more slowly than feedforward commands. This increased reliance on feedback can therefore slow the speech system and make the production of rapid sequences of sounds more vulnerable to disruption.

Overview of the DIVA model

Evidence supporting these mechanisms comes from neuroimaging studies that have identified differences in activity across several components of the speech motor network in individuals who stutter, including premotor cortex regions, sensorimotor speech areas, cerebellar circuits, and midbrain structures associated with basal ganglia pathways (Watkins et al., 2008).  These findings suggest that stuttering may involve differences in how neural signals are coordinated across the broader speech motor network rather than abnormalities confined to a single brain region.

All of  these findings suggest that stuttering may arise when the neural systems responsible for initiating and sequencing speech motor programs become less stable. Rather than reflecting anxiety or low self confidence, stuttering appears to involve differences in how the brain coordinates the rapid sequence of motor commands required for fluent speech.

Further Reading

Hickok, G., Houde, J., & Rong, F. (2011). Sensorimotor integration in speech processing: Computational basis and neural organization. Neuron, 69(3), 407–422.

Bouchard, K. E., Mesgarani, N., Johnson, K., & Chang, E. F. (2013). Functional organization of human sensorimotor cortex for speech articulation. Nature, 495(7441), 327–332.

References

Alm, P. A. (2004). Stuttering and the basal ganglia circuits: A critical review of possible relations. Journal of Communication Disorders, 37(4), 325–369. https://doi.org/10.1016/j.jcomdis.2004.03.001

Chang, S.-E., & Guenther, F. H. (2020). Involvement of the cortico–basal ganglia–thalamocortical loop in developmental stuttering. Frontiers in Psychology, 10, 3088. https://doi.org/10.3389/fpsyg.2019.03088

Neef, N. E., & Chang, S.-E. (2024). Knowns and unknowns about the neurobiology of stuttering. PLOS Biology, 22(2), e3002492. https://doi.org/10.1371/journal.pbio.3002492

Tourville, J. A., & Guenther, F. H. (2011). The DIVA model: A neural theory of speech acquisition and production. Language and Cognitive Processes, 26(7), 952–981. https://doi.org/10.1080/01690960903498424

Watkins, K. E., Smith, S. M., Davis, S., & Howell, P. (2008). Structural and functional abnormalities of the motor system in developmental stuttering. Brain, 131(1), 50–59. https://doi.org/10.1093/brain/awm241