Stuttering triggers cortisol. Cortisol degrades the motor circuit. The circuit fails again. FluentPlay breaks that loop — personalized, voice-driven games that lower your cortisol response during sessions, weaken the conditioned fear over time, and prepare you to speak in the real world with less fear. No drugs. No clinic required.
When someone stutters, the problem isn't in their mouth — it's in the brain's timing system. The part of the brain that plans speech sends signals that are mistimed or unstable. FluentPlay's games are designed around this science: they track how speech is produced, syllable by syllable, and show where the timing breaks down.
Technical: the speech-motor circuit — from phonological encoding in the IFG to motor gating in the putamen to execution at M1 — breaks down at specific points. FluentPlay's PAD framework monitors those breakdowns in real time, per syllable, during gameplay.
FluentPlay models stuttering as predictable instability in speech-motor timing during pre-articulatory planning — not at articulation onset. The breakdown happens before your mouth moves. Every factor is measurable and trainable.
Pressure isn't about phonetic complexity. It's about how much neural load a specific word carries for you, based on your history with it.
A word like "cat" is phonetically simple — one syllable, common consonants. But if you've stuttered on "cat" hundreds of times, your brain has hardened a cortisol-linked stress response around that exact motor sequence. The pre-SMA fires the plan, but the basal ganglia gates it with accumulated hesitation. The wiring remembers.
Pressure is measured per syllable through voice onset time, formant stability, and F0 variability. High Pressure on a simple word is the signature of learned motor disruption.
Anxiety is the cortisol spike that fires between knowing what you want to say and saying it. The amygdala detects the feared word approaching and triggers cortisol release before the motor plan even reaches execution.
This isn't general nervousness — it's word-specific anticipatory cortisol. You can be completely relaxed in a conversation and still feel a spike of dread when you see a specific sound coming. That cortisol surge modulates the basal ganglia gating system, making the putamen's timing more erratic.
Anxiety is independent of Pressure but amplifies it. A high-Pressure word with low Anxiety may flow fine. Add the anticipatory cortisol spike and the gate misfires.
Disfluency isn't a score you control — it's a signal that emerges. It's the product of Pressure and Anxiety interacting, not the sum. This distinction matters.
If Pressure is 5 and Anxiety is 2, the output isn't 7 — it's 10. If Pressure is 20 and Anxiety is 20, the output isn't 40 — it's 400. The multiplicative relationship means that reducing either factor dramatically reduces the output. Drop Anxiety in half and you cut Disfluency in half, even if Pressure stays the same.
This is why exposure therapy works on Pressure while real-time feedback stabilizes both. You don't need to fix everything — you need to break the multiplier.
Ground measures whether cortisol is settling or accumulating across a session. It tracks the trajectory — are Pressure and Anxiety trending down as you practice, or is the stress response building up?
A rising Ground score means the cortisol response is habituating. The basal ganglia gates are becoming more reliable with each repetition as the neurochemical load decreases. A falling Ground score means cortisol is accumulating — you may need to step back to an easier phrase or take a break.
Ground is the factor that tells you whether the cortisol loop is weakening. Not on a single word, but across the whole session. It's the foundation for progressive difficulty and real-world transfer.
Pressure and Anxiety are independent inputs. Disfluency emerges from their interaction — it's not additive, it's multiplicative. Ground measures whether your system is settling or accumulating strain.
Pick a word below. Watch what happens in your brain when it's feared — and how repeated practice rewires the cortisol response over sessions. The goal is not just fluency in the game. It's lowering your cortisol baseline so that speech techniques carry into real conversations with less fear.
Phonetic complexity doesn't predict stuttering. Neural history does. Tap each word to see the difference.
Select a word below to see where the circuit breaks
Select a word above to see why its PAD scores look the way they do — and where the signal breaks down in the speech-motor circuit.
Each game listens to your speech through the microphone and responds in real time. You don't type or tap — you talk. The games are designed to exercise specific parts of how your brain plans and produces speech — and they personalize to you. Choose your own phrases, target your hardest sounds, and build sessions around the words that matter most.
Easy onset and prolonged speech are among the most widely used fluency shaping techniques. Rainbow Syllables operationalizes these by requiring sustained, controlled voicing across an entire phrase. The volume zone feedback reinforces proprioceptive awareness — the speaker learns to feel and regulate their airflow and laryngeal tension in real time, which is the foundation of easy onset work.
Phrase-level production exercises the full basal ganglia-thalamocortical loop. Each syllable requires the putamen to gate the next motor plan in sequence. By practicing full phrases with real-time feedback, the speaker is training the SMA to coordinate smoother handoffs from one articulatory gesture to the next — the exact point where stuttering disrupts the timing chain.
Continuous phonation and coarticulation drills are standard in fluency therapy. Sound Bridge isolates the exact moment where voicing breaks down — the transition between two sounds. SLPs use connected speech tasks to train this, but most tools don't give real-time feedback on whether the speaker actually sustained voicing through the boundary. Sound Bridge does.
Coarticulation is controlled by the premotor cortex, which programs the specific sequence of articulatory muscle movements. In stuttering, premotor programming is less reliable during transitions — the abstract phonological plan fails to convert smoothly into continuous motor commands. Sound Bridge directly trains this conversion by requiring unbroken voicing across phoneme boundaries, strengthening the feedforward control described in the DIVA model of speech production.
Avoidance reduction therapy (Sheehan) and voluntary stuttering (Van Riper) are foundational approaches in stuttering treatment. The principle: avoidance reinforces fear, and fear drives more avoidance. Summit applies structured exposure — the speaker confronts their most feared words through repetition until the emotional charge diminishes. This is the same habituation mechanism used in cognitive behavioral approaches to anxiety.
Word-specific fear activates the amygdala, which modulates the basal ganglia gating system. When a feared word triggers an anxiety response, the putamen's motor gating becomes more unreliable — the gate misfires, producing blocks and repetitions. Repeated voluntary production of the feared word reduces the amygdala's threat response over time, which stabilizes the downstream motor gating circuit. Summit quantifies this process: the progress bar is a visible record of exposure volume.
Motor learning principles — specificity of practice, distributed practice, and variable practice — are the foundation of speech-motor training. Phoneme Drill isolates individual sound production so the speaker can build accuracy and consistency at the smallest unit before combining sounds into connected speech. Rhythm Pad trains proprioceptive control (hitting volume targets). Cadence trains temporal regulation (steady rhythm). Bubble Hunt trains precision under cognitive load (accuracy with time pressure).
Isolated phoneme production targets the M1 orofacial region — the final cortical output for speech motor commands. By training individual sound production with immediate feedback, the speaker strengthens the mapping between motor plan and motor execution. The cerebellum provides the millisecond-level timing coordination required here. The three sub-modes exercise different aspects of cerebellar-cortical coordination: volume regulation, temporal pacing, and adaptive motor control under varying demands.
Six games are live now with early-access partners. The suite is expanding. Drop your email to request access.
Will Carbone is the founder of FluentPlay Technologies. He's stuttered since childhood — the kind that teaches you to plan every sentence before you say it.
He came to speech technology from synthetic mRNA manufacturing — process development, tech transfer, production science. When he looked at what existed for people who stutter, he found clinical tools that hadn't evolved in decades and apps disconnected from the neuroscience. The research was clear that stuttering is a motor timing problem. The tools ignored it.
He left biotech, learned to build, and made the tools himself. Voice-driven games. Real-time speech scoring. A detection framework grounded in how the speech-motor circuit actually works. SLP and neuroimaging partnerships to keep it honest.
FluentPlay operates out of Somerville, Massachusetts. The games work because they were built by someone who needs them.
Have a question? Want to try the games with your SLP? Looking for tools for your child? Just want to say hi? Will reads every message personally.