TL;DR. DAF research dates from 1950 (Bernard Lee). The immediate effect — a substantial reduction in stuttering frequency during use — is one of the more replicable findings in stuttering science. Modern fMRI work shows that DAF normalises atypical brain activity patterns associated with stuttering. Long-term carryover is real but more variable than the immediate effect. The strongest evidence supports DAF as an effective component of a broader practice routine.
A short history
The DAF effect was discovered by accident. In 1950–1951, Bernard Lee was working on military and aviation communications when he noticed that delaying a speaker's own voice playback caused stuttering-like disfluency in fluent speakers. The phenomenon was named the Lee effect and initially studied as a disruption of normal speech.
Within a few years, researchers including Goldiamond noticed the inverse: people who stutter became more fluent under DAF. The first formal therapeutic application followed in the 1960s, with delays in the 100–250 ms range. The technology was bulky — large reel-to-reel devices. Effects were strong enough that by the 1970s DAF was a recognised, if underused, fluency-shaping technique.
The 1990s saw three major shifts:
- Miniaturisation. Hardware DAF devices became wearable.
- Theoretical work. Joseph Kalinowski and colleagues at East Carolina University conducted a long series of studies refining the understanding of altered auditory feedback (DAF + FAF + masking).
- Smartphone and software. By the late 2000s, software DAF on consumer phones brought the technique to mass-market practicality.
The Kalinowski review period
The most-cited synthesis of DAF research comes from Joseph Kalinowski and Tim Saltuklaroglu in the early 2000s. Their reviews — particularly the 2003 work consolidating altered auditory feedback findings — established the core empirical picture:
- DAF produces an immediate, robust reduction in stuttering frequency.
- The effect is largely independent of severity — both mild and severe stuttering respond.
- The effect holds across reading and spontaneous speech, with reading typically showing slightly larger effects.
- Optimal delay clusters around 50–200 ms, with substantial individual variation.
- DAF effects are preserved when frequency-altered feedback is added (FAF + DAF combinations).
- Effects are present in first-language speakers across multiple languages.
A representative effect size: in a reading task, an adult who normally stutters on 8% of words might stutter on 1.5–3% under DAF — a 60–80% reduction. That reduction is clinically meaningful in reading studies, but it is not a guarantee for spontaneous speech or long-term carryover.
What modern fMRI shows
Brain imaging of stuttering has grown rapidly over the past 20 years. The findings most relevant to DAF:
- Atypical activity patterns in basal-ganglia–cortical timing circuits during stuttered speech.
- Reduced left inferior frontal gyrus activity — part of Broca's area — during speech planning in people who stutter.
- Compensatory right hemisphere activity during stuttered speech.
- Task-related shifts under choral and DAF conditions — some studies report changes in right-hemisphere recruitment and basal-ganglia patterns toward fluent-speech task patterns.
A 2020 review in Frontiers in Neuroscience and several PMC-indexed studies discuss DAF in relation to auditory tracking and compensatory networks that may reduce the impact of stuttering-related timing differences during use.
The brain isn't being "tricked" by DAF; it's being given a clearer signal to track, and it tracks it.
Long-term outcomes — where the evidence softens
The immediate effect is strong. The long-term effect is more variable.
Studies of long-term DAF use — does the benefit persist once the device comes off — fall into three patterns:
- Some users carry over substantial gains. Their non-DAF speech improves measurably after weeks of regular DAF practice.
- Some users carry over partial gains. The technique they practised under DAF (slower pace, gentler onset) persists even without the audio.
- Some users show effect only during use. Stuttering returns when the device is off, with no persistent improvement.
Predictors of who carries over are not fully understood. The practical takeaways from current studies:
- DAF works best as part of a structured practice routine with technique work, not as a passive, all-day device.
- The presence of transfer practice — short blocks of speech without DAF — appears to help carryover.
- Combining DAF with therapy from a stuttering-specialist SLP appears more reliable than using DAF as a passive, stand-alone tool.
What about the negative findings?
Not every study is positive. Where DAF tends to fall short:
- A small but consistent percentage of users (~10–20%) find DAF unhelpful, distracting, or aversive.
- Studies with very short delays (under 30 ms) show smaller or absent effects.
- Studies with hardware that has unstable latency (early Bluetooth, mismatched hardware-software combos) show degraded effects.
- Studies with users who have severe co-occurring conditions (significant hearing loss, neurogenic stuttering, severe psychogenic stuttering) show variable response.
The evidence base is positive on balance, not unanimous. That's typical for any non-pharmaceutical intervention, and it should temper but not undermine confidence in DAF as a tool.
Comparison to other interventions
How DAF compares to other behavioural interventions for stuttering:
- vs fluency-shaping techniques alone — DAF often works with fluency shaping, not instead of it. The combination is stronger than either alone.
- vs medication — there is no drug approved specifically to treat stuttering, and the medications sometimes used off-label carry systemic side effects; DAF has none.
- vs intensive program-based therapy — Lidcombe (children) and Camperdown (adults) have published clinical-trial evidence. DAF complements these programs but does not replace them.
Open questions
Things the research has not yet fully answered:
- Why some users carry over and others don't. Predictors are partial.
- Optimal delay individualisation. Best-delay-for-this-person research is thin.
- Effects in children. Most DAF research is on adults; child effects are less well characterised, and use in children is generally cautious.
- DAF + neuromodulation. Early work combining DAF with TMS or tDCS is preliminary.
- Effects on covert stuttering. Most research measures audible stuttering frequency; effects on internal struggle and avoidance are less measured.
How to think about the evidence
If you're a person who stutters considering DAF, the practical takeaways:
- The immediate effect is well-supported. Many adults notice less stuttering during early sessions, though individual response varies.
- Long-term benefit is real but practice-dependent. Use DAF as a training tool with technique work, not as a standalone support.
- Combine it with therapy when possible. A few sessions with a stuttering-specialist SLP plus daily DAF practice is better supported than passive app use alone.
- Get the boring parts right: wired headphones, a defined delay, and structured time blocks. These details are what separate effective use from ineffective use.
For the mechanism, see how DAF works. For the practical tool, see what is DAF and exercises.
Practise for 5-15 minutes with wired headphones. StutterFlow is a practice tool, not a cure or a replacement for speech therapy.
