Audiomotor prediction errors drive speech adaptation even in the absence of overt movement

This is an uncorrected proof. Figures Abstract Observed outcomes of our movements sometimes differ from our expectations. These sensory prediction errors recalibrate the brain’s internal models for motor control, reflected in alterations to subsequent movements that counteract these errors (motor adaptation). While leading theories suggest that all forms of motor adaptation are driven by learning from sensory prediction errors, dominant models of speech adaptation argue that adaptation results from integrating time-advanced copies of corrective feedback commands into feedforward motor programs. Here, we tested these competing theories of speech adaptation by inducing planned, but not executed, speech. Human speakers were prompted to speak a word and, on a subset of trials, were rapidly cued to withhold the prompted speech. On standard trials, speakers were exposed to real-time playback of their own speech with an auditory perturbation of the first formant to induce single-trial speech adaptation. Speakers experienced a similar sensory error on movement cancellation trials, hearing a perturbation applied to a recording of their speech from a previous trial at the time they would have spoken. Speakers adapted to auditory prediction errors in both contexts, altering the spectral content of spoken vowels to counteract formant perturbations even when no actual produced speech coincided with the perturbed feedback. Such adaptation was not observed when participants passively listened to perturbed feedback without the intention to speak, ruling out observational learning as the cause of adaptation in movement cancellation trials. These results suggest that prediction errors, rather than corrective motor commands, drive audiomotor adaptation in speech, building on recent findings in reaching. Citation: Parrell B, Bae M, Naber C, Kim OA, Niziolek CA, McDougle SD (2026) Audiomotor prediction errors drive speech adaptation even in the absence of overt movement. PLoS Biol 24(6): e3003862. https://doi.org/10.1371/journal.pbio.3003862 Academic Editor: Johanna Rimmele, Max-Planck-Institut for Empirical aesthetics, GERMANY Received: January 30, 2026; Accepted: June 4, 2026; Published: June 15, 2026 Copyright: © 2026 Parrell et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: All data and analysis scripts are available at https://osf.io/xqztr. Funding: This work was supported by grants R01 NS134754 (C.A.N, B.P, & S.D.M.), R01 DC019134 (C.A.N. & B.P.), R01 DC017091 (B.P.), and R01 NS132926 (S.D.M.) from the National Institutes of Health and BCS 2120506 (C.A.N. & B.P.) from the National Science Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. Abbreviation:: LPC, linear predictive coding Introduction Implicit motor adaptation is a prime example of the brain’s predictive capacities. In the standard framework, motor adaptation occurs when the predicted sensory consequences of a descending motor command are compared against feedback, and an internal model of those motor commands is adjusted to reduce observed discrepancies (i.e., sensory prediction errors) [1– 3]. One lingering question in this framework is what serves as the input signal for the generation of sensory predictions that drive motor learning. In some models, that input signal is an efference copy generated from the descending motor command, which is then used to predict future states [4– 7]. In other work, the input signal is defined as a higher-level motor plan [8– 12]. It remains an open question whether motor planning or execution are sufficient for the computation of sensory predictions and errors (Fig 1), and whether there are common requirements for adaptation across motor and sensory subsystems. Note the diagram uses the following standard conventions: (1) information that is sent to multiple locations is shown by multiple arrowheads splitting from a single origin; (2) system updates are represented by an arrow crossing behind a box; (3) arrows that represent different flows of information that need to cross for visual display purposes are represented with a bump in one of the two lines. In both theories, reafferent feedback about movement outcomes is compared against sensory predictions generated by forward models. Any discrepancies between observed and predicted sensory outcomes generate a sensory prediction error. A: In this model, sensory prediction errors drive adaptation indirectly: they are used by the motor controller to update motor commands (sensory feedback control), and these feedback-based motor commands are then used to adapt the motor control system (arrow labeled Implicit Adaptation). This is the basic mechanism of adaptation in the DIVA model of speech production [14]. When movement is canceled (pink box and arrows), this model predicts adaptation should not occur, as no feedback-based motor commands will be generated. B: In this model, sensory prediction errors directly drive adaptation of the motor control system (arrow labeled Implicit Adaptation). As adaptation and (feedback-based) motor control are separated, this model predicts adaptation should occur even in the absence of motor movement as long as sensory prediction errors are still generated. The computations underlying adaptation in speech production are a point of particular debate, with competing models suggesting either that ongoing feedback during actual execution of speech commands drives adaptation to audiomotor prediction errors by incorporating feedback-based motor commands into future movement plans [13, 14] or, conversely, that adaptation instead relies directly on auditory errors generated by the comparison of auditory reafference with auditory predictions deriving from a task or planning-level controller rather than by motor efference [15, 16]. To date, it has been difficult to distinguish between these competing theories as they make similar predictions in existing speech adaptation paradigms. Resolving this issue has implications for understanding, and potentially treating, the mechanisms underlying impaired speech adaptation in neural disorders such as ataxia [17, 18] as well as changes that occur in speech in the presence of sensory impairments, such as deafness, that degrade sensory feedback [19– 22]. Moreover, addressing this question will aid our understanding of how analogous adaptation processes may differ or not across different motor and sensory systems. In upper-limb control, several lines of evidence support the notion that “upstream” motor planning is sufficient to drive adaptation. One piece of evidence for this is the discovery that internal signals preceding a movement, such as planning “lead-in” and “follow-through” motions, can partition otherwise catastrophically interfering motor memories [9, 10]. Indeed, a range of similar contextual pre-cueing effects have been observed in motor adaptation, supporting the idea that preparatory signals seed predictive computations within the adaptation system [12, 23– 27]. However, there is no comparable direct evidence for speech adaptation and, even for reaching, these studies could not rule out the potential necessity of motor-based efference copy. We have recently developed an experimental paradigm in limb control that is capable of teasing apart plan- versus motor-based prediction error in sensorimotor adaptation, which we used to establish that implicit upper-limb motor adaptation could be induced in response to sensory errors that were unaccompanied by movements [28]. Human subjects performed a “Go/NoGo” task that involved planning, but occasionally not executing, reaches to visual targets. On certain critical trials, subjects successfully inhibited reach commands but still observed a “virtual” sensory prediction error delivered via a visual cursor rotated relative to the planned reaching direction. These trials induced adaptation in subsequent movements, even though no movement had accompanied the error. In our view, these results offer particularly compelling support for the aforementioned plan-based model of adaptation. Here, we applied this design to speech adaptation in order to answer one of the major outstanding questions about sensorimotor adaptation in speech: whether ongoing feedback from executed speech movements is necessary for adaptation or whether speech adaptation can leverage sensory predictions based only on planned speech acts. Based on our previous work in limb control, we expect to observe adaptation when speech movements are planned, but not executed, when paired with “virtual” auditory sensory errors. Such a result would strongly support models of speech adaptation driven directly by sensory prediction errors over models where adaptation is only driven through feedback-based motor commands. Moreover, by expanding this paradigm from a laboratory reaching task to speech, this work provides a critical test of the domain-generality of our previous findings. Speech differs in many ways from reaching, including sensory feedback modalities (auditory versus visual feedback), the neural basis of control (bulbar versus spinal control), and the ultimate task goal (communication versus interacting with the physical environment). If, despite these substantial differences, similar results are found in speech and reaching, this would suggest a shared neurocomputational mechanism driving sensorimotor adaptation. Results We implemented a task that allowed us to measure speech adaptation both with and without movement execution during perceived auditory prediction errors (Fig 2). Trials were organized into three-trial “triplets”, where the middle trial received a perturbation to the first vowel formant (F1), one of the resonances of the vocal tract determined by the position of the tongue, lips, and jaw which distinguish vowels from one another (Fig 2A). On “movement” trials, participants produced speech when cued by a visual stimulus. To measure adaptation in the absence of movement execution, we had participants occasionally inhibit a cued speech movement while still presenting them with a well-timed audiomotor prediction error (“no-movement” trials), using a formant perturbation applied to a recording of the speech produced on the immediately preceding trial. We measured adaptation for the movement and no-movement trials as the change in F1 from the trial preceding the perturbed trial to the two trials immediately following the perturbed trial (adaptation and aftereffects, Fig 2E). A: Illustration of the relationship between tongue position and vowel formants. The first two vowel formants (F1, F2) serve to distinguish between different vowels. B: Schematic of movement and no-movement trials. In movement trials, a stimulus word appeared on the screen and participants read the word out loud. Participants heard their own voice, with the first vowel formant (F1) perturbed or unperturbed. In no-movement trials, the stimulus word turned red 200 ms after appearing. Participants were instructed not to produce speech on these trials. Instead, the audio recording from the previous trial was played back to the participants with F1 perturbed. C: Triplet design. In each triplet, trials 1 and 3 were always movement trials with no perturbation applied. Trial 2 was variable—a movement or no-movement trial—and always had a perturbation applied. D: Experiment structure. The experiment consisted of 36 blocks which alternated between three stimulus words. Each block consisted of 6 triplets (2 with movement middle trials, 4 with no-movement middle trials) and 3 distractors (always movement trials) to disrupt the rhythm of the perturbations. E: Data analysis. Each trial shows a spectrogram of an example speech trial; cyan lines show produced vowel formants (F1 and F2) and blue lines show formants in headphone signal. For each trial, the produced F1 (the lowest frequency formant) was averaged between 50-125 ms after vowel onset (green shaded portion) to avoid coarticulatory effects of the initial consonant in each stimulus word and potential online compensatory effects. Single-trial adaptation was measured as the change in F1 from trial 1 to trial 3 within each triplet. Aftereffects were measured as the change from trial 1 of one triplet to the trial immediately following trial 3 of that triplet (either a distractor or trial 1 of the next triplet). Despite the instruction to inhibit speech, participants nonetheless produced some acoustic speech material on a subset of no-movement trials (average of 22%, range 3%–62%, Fig 2A). The prevalence of these inhibition failures suggests that our method was successful in driving participants to plan speech at the onset of the stimulus in these no-movement trials. In order to isolate learning in the absence of movement, only no-movement trials where participants successfully inhibited overt speech production were used to measure adaptation. After exclusion of trials where participants failed to inhibit overt speech as well as trials with errors in production (participants produced the wrong stimulus, or unresolvable errors in formant tracking), the analyzed dataset included, for each participant, an average of 71 movement triplets (64–72 across participants) and 111 no-movement triplets (52–140 across participants). Because the expected effect size of single-trial adaptation (~3 mels in Hantzsch and colleagues [29], Fig 3A) is much smaller than the expected standard deviation of F1 in typical speech (~10–15 mels), our primary outcome measurement combined measurements taken from the trial immediately following the perturbation (adaptation) as well as the subsequent trial (aftereffects), in order to increase the power of our analysis and reduce noise. Participants produced robust single-trial learning following perturbations to both movement and no-movement trials (Fig 3A and 3B, main effect of perturbation direction F(1,10,724) = 59.25, p < 1 × 10−13, partial R2 = 0.005). For movement trials, F1 following trials with a downward F1 perturbation was significantly higher (2.3 ± 0.8 mels) than following trials with an upward F1 perturbation (−6.2 ± 0.8 mels, t(10722) = 7.5, p < 0.0001, d = 0.23). Single-trial adaptation on movement trials was 3–4 times larger than in our previous work [29], likely reflecting better control of stimuli in the current paradigm (see Discussion). Underlying data for all panels is available in file EXP1_statTable_30s_50-125.txt at https://osf.io/xqztr/. A: The percentage of no-movement trials on which participants failed to inhibit vocal production. The average failure rate was 22% across participants. These trials were excluded from further analysis. B: F1 adaptation index (see Methods) following trials with downward and upward perturbations applied in movement (left) and no-movement (right) conditions. Estimated marginal means and standard errors shown in green and magenta. Gray indicates single-trial learning observed in Hantzsch and colleagues [29]. C: As for (B), showing data for individual participants. Means and standard error across participants are shown in dark, solid colors. Individual participant means shown as lighter colored dots connected by gray lines. D: Relationship across participants between the magnitude of differential adaptation (responses following downward perturbation minus responses following upward perturbations) in movement and no-movement conditions. ** p < 0.01, **** p < 0.0001. Crucially, following no-movement trials, F1 was significantly higher in trials following downward perturbations (0.5 ± 0.7 mels) compared to upward perturbations (−2.1 ± 0.8 mels, t(10736) = 2.9, p = 0.004, d = 0.07). This result supported our hypothesis that speech adaptation could proceed without actual movement execution. The magnitude of adaptation was smaller in no-movement trials compared to movement trials, reflected by a significant interaction between movement condition and perturbation direction (F(1,10,725) = 16.3414, p < 0.0001, partial R2 = 0.002). Overall, the magnitude of adaptation in no-movement trials was roughly one-third of that found in movement trials, similar to our previous results in adaptation of upper limb movements [28]. (Potential reasons for this replicated difference in effect size are addressed in the Discussion.) We then looked at a between-subjects correlation between adaptation effect sizes on movement versus no-movement triplets. Indeed, there was a moderate, though statistically marginal, positive relationship between learning in movement and no-movement trials across participants in the expected direction (Fig 3D, r = 0.36, p = 0.053). As an additional analysis, we considered single-trial adaptation (in trials immediately following the perturbation, t + 1) and aftereffects (in the next following trial, t + 2) separately (Fig 4). The results for both adaptation and aftereffects were consistent with the combined results. For adaptation, there was a significant main effect of perturbation direction (F(1,5362.3) = 32.7, p < 1 × 10−7, partial R2 = 0.006) as well as a significant interaction between perturbation direction and movement condition (F(1,5362.3) = 10.2, p = 0.001, partial R2 = 0.002). The difference in F1 between upward and downward trials was significant for both movement (8.6 ± 1.5 mels, t(5363) = 5.7, p < 0.0001, d = 0.24) and no-movement conditions (2.5 ± 1.2 mels, t(5369) = 2.014, p = 0.04, d = 0.07), though the magnitude of this difference was again smaller for the no-movement condition. Underlying data for all panels is available in file EXP1_statTable_30s_50-125.txt at https://osf.io/xqztr/. A: Differential adaptation shown separately for trials immediately following the perturbation (+1, adaptation) and the next trial (+2, aftereffects) in the movement (left) and no-movement (right) conditions. Estimated marginal means and standard errors shown in green and magenta. Gray indicates single-trial learning observed in Hantzsch and colleagues [29]. B: As for (A), showing data for individual participants. Means and standard error across participants are shown in dark, solid colors while individual participant means are shown as lighter colored dots connected by gray lines. * p < 0.05; ** p < 0.01. The results for aftereffects were similar: Upward and downward perturbations led to differences in F1 in both movement (8.5 ± 1.7 mels, t(5328) = 4.997, p < .0001, d = 0.22) and no-movement (2.9 ± 1.4 mels, t(5341) = 2.110, p = 0.03, d = 0.07) conditions (main effect of perturbation direction: F(1,5,355) = 27.2, p < 1 x 10−6, partial R2 = 0.005); and the magnitude of this difference was smaller in the no-movement condition (interaction: F(1,5,355) = 6.6, p = 0.01, partial R2 = 0.001). One possible explanation for the adaptation observed in no-movement trials is that participants might have learned from passive exposure to the auditory feedback error without necessarily generating a sensory prediction. In certain contexts, such observational learning has been observed in limb control [30, 31]. To test whether the adaptation observed in no-movement trials was a product of observational learning, we conducted a second study (Experiment 2) which mirrored the design of the initial study (Experiment 1), but substituted “observation” trials for no-movement trials. In this experiment, visual cues indicating whether to speak (movement trials) or remain silent (observation trials) were presented prior to stimulus onset, so that motor planning (and concurrent sensory prediction) would not occur in observation trials (Fig 5A). After exclusion of trials with errors in production, the analyzed dataset included, for each participant, an average of 70 movement triplets (63–72 across participants) and 137 no-movement triplets (127–140 across participants). Underlying data for all panels is available in file EXP2_statTable_30s_50-125.txt at https://osf.io/xqztr/. A: Schematic of movement (left) and observation (right) trials. In movement trials, a visual cue to speak appeared before a stimulus word, and participants read the word aloud. As in Experiment 1, participants heard their own voice with either perturbed or unperturbed F1 feedback. In observation trials, a cue to not speak was shown before a stimulus word, and participants heard the F1-perturbed audio recording from the previous trial. B: Estimated marginal means and standard errors of the F1 adaptation index (see Methods) following trials with downward and upward perturbations in the movement (left) and observation (right) conditions. Green represents the F1 adaptation index observed in movement trials in Experiment 1. C: As for (B), showing data for individual participants. Means and standard errors across participants are shown in dark, solid colors. Individual participant means shown as lighter colored dots connected by gray lines. D: Experimental results by trial number relative to the middle trial of each triplet. Estimated marginal means and standard errors of differential adaptation for the trial immediately following perturbation (+1, adaptation) and the subsequent trial (+2, aftereffects) in the movement (left) and observation (right) conditions. Green indicates corresponding results from Experiment 1 movement trials. E: As for (D), showing data for individual participants. Means and standard errors across participants are shown in dark, solid colors while individual participant means are shown as lighter colored dots connected by gray lines. **** p < 0.001. We compared F1 changes following perturbations between movement and observation trials. Critically, adaptation was observed only in the movement condition (Fig 5B and 5C). In movement trials, F1 was significantly higher following downward perturbations (1.16 ± 0.75 mels) than following upward perturbations (−6.93 ± 0.75 mels; t(11824) = 7.84, p < .0001, d = 0.24). The effect seen here was a close replication of the movement condition in Experiment 1. In contrast, no significant difference was found in the observation condition (downward: −0.80 ± 0.56 mels; upward: −1.28 ± 0.55 mels; t(11846) = 0.63, p = .529). The contrast was reflected in a significant interaction between perturbation direction and movement condition (F(1,11,842) = 35.27, p < .001, partial R2 = 0.003), suggesting that the adaptation observed in no-movement trials in Experiment 1 cannot be explained by observational learning alone. Separate analyses of adaptation and aftereffects trials each show the same pattern as the combined measure (Fig 5D and 5E). In movement trials, significant learning was observed in adaptation trials (t(5910) = 5.29, p < .0001, d = 0.23), as well as aftereffects trials (t(5882) = 5.79, p ≤ .0001, d = 0.256). In contrast, there was no evidence of learning in observation trials, neither in adaptation (t(5933) = −0.38, p = .702) nor aftereffects (t(5907) = 1.24, p = .216). As for the combined measure, the interaction between perturbation direction and condition was significant for both adaptation (F(1,5,919) = 20.09, p < .0001, partial R2 = 0.003) and aftereffects (F(1,5,909) = 15.47, p < .0001, partial R2 = 0.003). Discussion Here, we demonstrated that speech adaptation can occur even when audiomotor errors are not accompanied by movement execution. First, across two experiments we replicated the result that when participants were exposed to a perturbation of their first vowel formant on a single trial, changes in speech articulator movements that opposed that perturbation were visible on the two trials following the perturbation. These results confirm that single-trial speech adaptation observed in previous paradigms designed to examine online compensatory movements [29] can be reliably elicited with our triplet design. Remarkably, adaptation occurred not only when participants heard an auditory perturbation during speech, but also when participants planned to produce speech but withheld overt speech production and instead heard playback of their own speech from a previous trial with the perturbation applied. That is, when participants putatively generated a movement plan and a concomitant auditory prediction, they adapted their movements to correct for observed audiomotor prediction errors even though they did not produce overt speech in conjunction with those errors. Conversely, when motor planning was absent, as in observation trials, adaptation did not occur, ruling out observational learning as a mechanism for the observed changes in the no-movement trials. These results extend our recent findings in upper-limb visuomotor adaptation [28] to a new and very different motor domain: speech differs from reaching in sensory feedback modality (audition versus vision), locus of neural control (bulbar versus spinal), and movement goals (communication versus environmental interaction). Despite these differences, the observed behavior in the two motor domains was strikingly similar. By dissociating observation from motor planning, the current study supports the idea that adaptation in speech, and likely more broadly across motor domains, involves a learning mechanism that depends on internal predictions generated during motor planning. Why we failed to observe observational learning in Experiment 2, given that such learning has been seen in upper-limb motor control [30, 31], remains an open question. We note that the precise conditions in the current study meaningfully differ from existing limb control studies on observational learning. For instance, a prior experiment that induced observational visuomotor reach adaptation [30] involved participant observation of the actual motor learning process, including corrective behaviors. In contrast, participants in the current study only observed altered sensory stimuli and were not allowed to observe an external learner. As another example, in Pawlowsky and colleagues [31], observational learning was seen when participants observed a sensory error (rotated cursor movement from a home position to a visually displayed target) while making an unrelated limb movement. While this is more similar to the current study in that no error-correction process was observed, it differs in whether or not participants were performing a motor action. Based on these distinctions, it may be that observing the learning process or engaging in an actual movement is critical for transferring observations to changes in action. Alternatively, the difference in results between our speech adaptation data and others’ limb movement data may lie in the nature of the movement targets themselves. While limb control studies have well-defined motor targets (e.g., visually displayed circles), speech targets are inherently individualized and flexible. In fact, speech has been shown to change by approximating the acoustic characteristics of perceived speech produced by other talkers (Goldinger, 1998; Shockley and colleagues, 2004; Babel, 2012; Babel and Bulatov, 2012). In the current experiment, it is unclear whether participants perceived the perturbed auditory feedback as an external speaker’s voice, which could drive phonetic convergence even in passive social situations (Goldinger, 1998; Namy and colleagues, 2002). If the altered feedback was perceived as another talker, it is possible that phonetic convergence may have been elicited in our study, thereby counteracting any observation-based adjustment and resulting in the non-significant net adaptation (though the reduction in convergence with a delay between the auditory stimulus and vocal productions seen in previous work [Goldinger, 1998] may make this explanation unlikely). In any case, the striking difference in behavior seen in no-movement trials, which elicit clear adaptation, and observation trials, which do not, strongly suggests that motor planning enhances audiomotor adaptation in speech, presumably by supporting the generation of a sensory prediction error, which is likely not present in observation trials. The current results suggest that audiomotor speech adaptation is driven, at least in part, by using sensory prediction errors to directly update internal models controlling speech articulators. Such updates are consistent with dominant models of sensorimotor learning in limb and oculomotor control [5, 32– 35] and with a recent model of speech motor learning [15]. However, these results are inconsistent with the dominant model of sensorimotor adaptation in speech production, which suggests that adaptation is driven by the incorporation of corrective movements generated by a feedback controller in response to sensory errors into future feedforward motor programs [13, 14, 36], which itself is partially based on other ideas in limb control [37, 38]. Notably, the magnitude of the change observed on the movement triplets (~8 mel difference between upward and downward perturbations) was roughly 2–3 times larger than we previously observed using the same analysis window [29]. Why did we see this large increase in effect size? One possibility is that our new design, which uses blocked repetition of the same stimuli, induces stronger learning signals than when stimuli are mixed across trials as in previous studies. This could be because adaptation in speech only partially generalizes to untrained words [39, 40], similar to the local spatial generalization of learning observed in reaching [41– 43]. Importantly, the larger effect size we observed in our movement trials may allow for more precise assays of factors that affect sensorimotor adaptation in speech, given that around 100 observations of learning can be obtained in the time it would typically take to obtain a single observation (e.g., mean asymptotic learning) using a more traditional paradigm with extended exposure to a repeated auditory perturbation. Intriguingly, the size of the reduction we observed in speech adaptation between movement and no-movement trials (roughly one-third, Figs 2 and 3) was very similar to the reduction previously observed in a similar reaching task [28], pointing to a domain-general explanation. We see several possible explanations for this observed reduction: First, it has been established, both in speech and reaching tasks, that temporal delays between movement and sensory feedback substantially impair adaptation to errors [44– 48]. In speech, the magnitude of adaptation is reduced by roughly 50% when auditory feedback is delayed by only 100 ms, and adaptation is essentially absent when delays reach 250–500 ms [49, 50]. This suggests that the critical comparison driving auditory error processing is highly temporally specific, particularly in speech. In our paradigm, it is impossible to know exactly when speech would have occurred on the no-movement trials; as an estimate, the latency of the speech feedback on a given no-movement trial was matched to the latency on a previous movement trial. This method is likely to have introduced variance between the timing of anticipated and perceived auditory feedback on no-movement trials, potentially leading to a reduction in the magnitude of adaptation. Second, reductions in somatosensory feedback inherent in no-movement trials may have suppressed the degree of adaptation possible. It has been suggested that, in limb control, adaptation may result not from a drive to reduce visual sensory error but rather from the recalibration of somatosensory and other sensory signals [51]. Consistent with this idea, putatively disruptive noninvasive stimulation to primary somatosensory cortex following visuomotor reach adaptation substantially decreases how much of this learning is retained [52]. Thus, when errors occur in the absence of somatosensory feedback, such as in our no-movement trials, it is possible that adaptation magnitude would be reduced. Finally, it may also be the case that adaptation in speech can be driven by auditory prediction errors both directly through updates to predictive internal models and indirectly through incorporation of previous feedback-based commands into feedforward motor programs. In fact, there is some support for this “dual input” idea in tasks requiring adapting to dynamic force-field perturbations during reaching [38]. Thus, an important note here is that while many results point to a plan-based sensory prediction error model of adaptation, both planning and ongoing feedback commands could both provide inputs that generate sensory predictions. Future work can directly assay if and how both sources of information fuel adaptation. Although direct neurophysiological data related to speech adaptation is limited, our results are consistent with invasive and non-invasive imaging studies that have examined sensory prediction in speech. When we speak, auditory cortical activity is suppressed relative to when we passively listen to the same sounds, a process thought to be driven by the cancellation of auditory reafference using predictive internal models [53– 56]. This suppression effect is reduced in less prototypical productions, suggesting that the prediction is based on a plan (or target) rather than motor efference [57– 59]. Recent work has shown that suppression is modulated during adaptation, and that the degree of modulation predicts the degree of learning, strongly suggesting that these predictions (and the resulting sensory prediction errors) play a critical role in adaptation [16]. Single-unit recordings in marmosets indicate that suppression of auditory cortex firing begins hundreds of milliseconds ahead of vocalization initiation [60], and stimulus-specific cortical modulation has been demonstrated in this pre-vocalization window in humans [61], suggesting that vocalization planning alone can modulate auditory cortex activity and may be sufficient to support error computation. Although we did not directly compare the magnitude of adaptation in response to the upwards and downwards perturbations, there is a clear asymmetry visible in the results, with responses to upward perturbations being much larger than responses to downward perturbations. The source of this asymmetry is not immediately apparent. However, similar asymmetries have been observed before, including in our previous study of single-trial adaptation across a large sample of speakers, though the direction of the asymmetry varies across studies and across vowels [29, 62– 64]. Notably, by focusing on the difference between adaptation to opposing formant perturbations, we were able to reliably detect learning despite this asymmetry in response magnitude. There are two caveats about our no-movement condition. First, although single-trial adaptation was robust in the no-movement condition, the overall effect was relatively small (~2 mels, d = 0.07). Though this magnitude is roughly equivalent to our previous work demonstrating single-trial adaptation (in movement trials) when stimulus words varied across trials [29], speech, in general, is substantially more variable than this, with a standard deviation in F1 of roughly 10–15 mels. Our triplet design, however, allowed for a high number of observations per participant in order to look for this modest effect. Second, while we excluded no-movement trials with any overt sound production of any kind, it is possible that participants nonetheless produced either articulatory movements without overt vocalization or subtle, subthreshold muscular activity on a subset of these trials. However, there is some evidence that, in response to a stop signal similar to that used here, inhibition of vowel-related articulatory movements and voicing most frequently co-occur [65]. This suggests that the acoustic signal used as a criteria for exclusion here is likely to also exclude most trials with overt vowel-related movement, though future work measuring muscle activity or movement more directly, such as with surface EMG of the masseter or articulatory tracking of the tongue, is needed to answer this question more definitively. Overall, our study suggests that the direct updating of internal models through sensory prediction errors can drive speech adaptation, though the reduced magnitude of adaptation in no-movement trials suggests other signals or timing constraints likely also play a role. In our view, speech planning may provide the critical sensory predictions that, when violated, lead to adaptation of internal models governing speech control. Critically, these results do not support models of speech adaptation that rely solely on the incorporation of feedback-based motor commands into future feedforward plans, suggesting that such models could be revised to incorporate direct updating of internal models through sensory error processing. Moreover, these results parallel our previous findings in upper-limb adaptation [28] in a novel motor domain (speech) and a different sensory system (audition). This commonality suggests that the algorithmic logic of motor adaptation shows parallels across domains; despite the many differences between speech and reaching, adaptation in both systems at least partially relies on a similar neurocomputational principle whereby sensory errors, based on plan-based predictions, drive updates to internal models. Methods Ethics statement Participants and power analyses Thirty participants were tested in each experiment (Experiment 1: 5 male/25 female, age range 18–45, mean age 23.4; Experiment 2: 4 male/26 female, age range 18–64, mean age 25.6). The sample size was determined using a bootstrapping procedure with effect sizes observed in our previous work [29], with the target of 90% power to detect a similar sized effect at α = 0.05. All participants were native speakers of American English, without any reported history of neurological, speech, or hearing disorders. All participants passed an automated Hughson-Westlake hearing screening (thresholds 25 dB HL or lower in both ears at 250, 500, 1,000, 2000, and 4,000 Hz). Task setup Participants were seated in a sound-insulated booth in front of a computer monitor. On each trial, a target word appeared on the screen in white text (Fig 1B). Each trial lasted 1.7 s from stimulus onset. Trials were separated by 1.25 s plus a random delay of 0–0.5 s. Participants were instructed to read the words as quickly as possible as they appeared on each trial. Two trial types were used—“movement” trials and “no-movement” trials, following our previous study in reach adaptation [28]. On the majority of trials (movement trials), the word stayed white for the duration of the trial. On a subset of trials (no-movement trials), the target word turned red 200 ms after it appeared and stayed red for the remainder of the trial. Participants were instructed to not produce overt speech if and when the target word turned red. 200 ms was chosen after pilot testing suggested that this delay results in the inhibition of speech on most, but not all, trials in the majority of participants. This delay was therefore long enough to elicit movement planning but short enough to enable mostly successful inhibition, thereby allowing us to test whether sensory feedback given in the absence of overt movement could drive speech adaptation. On movement trials, participants’ speech was recorded (AKG C520), digitized with a USB sound card (Focusrite Scarlett 2i2), processed through the Audapter software package [66, 67], and played back to the participants over closed-back, over-the-ear headphones (Beyerdynamic DT 770). Speech was played back at a volume of approximately 83 dB SPL and mixed with speech-shaped noise at approximately 60 dB SPL. The final level of the playback speech signal varied with the amplitude of participants’ produced speech. The noise, combined with the closed-back headphones, served to minimize potential perception of the participants’ own unaltered speech, which may have otherwise been perceptible through air or bone conduction. The latency of audio playback on our system is ~ 18 ms, as measured using the protocol suggested by [68]. Trials were organized into “triplets.” The first and last trials of each triplet were always movement trials with veridical auditory feedback (Fig 1C). On the middle trial of each triplet, the auditory feedback participants received was always perturbed, such that the first vowel formant (F1, Fig 1A) was either raised or lowered by 125 mels (a perceptually-calibrated measurement of frequency) throughout the trial using Audapter. Middle trials were either movement trials (⅓ of triplets) or no-movement trials with a stop signal (⅔ of triplets). On perturbed movement trials, the auditory feedback was perturbed in real time by identifying the vowel formants using linear predictive coding (LPC) and filtering the speech signal to introduce a shift to those formants. On perturbed no-movement trials, Audapter was used offline to apply the same shift to a recording of the participants’ production from the previous trial (i.e., the first trial of the triplet, which was always an unperturbed “movement” trial). On these trials, Audapter was initiated at the same time as on movement trials, but played back this perturbed speech signal rather than playing back the speech recorded in real time. Thus, the latency of the auditory feedback on each no-movement trial was the same as on the immediately preceding trial. Triplets were organized into blocks (Fig 1D). Each block contained 2 movement triplets and 4 no-movement triplets. Upward and downward frequency perturbations were balanced within each block. The order of triplets within each block was randomized. Each block additionally contained 3 distractor trials, randomly inserted between triplets, in order to disrupt the rhythm of the perturbations across trials. All distractor trials were movement trials with veridical auditory feedback. Each block used a single stimulus word (“bed,” “dead,” or “Ted,” all sharing the same target vowel/ɛ/); stimuli were pseudo-randomized across blocks such that there were an equal number of blocks for all stimuli, and no consecutive blocks used the same stimulus. The experiment consisted of 36 total blocks, yielding 216 total triplets. This resulted in 36 perturbations in movement blocks and 72 perturbations in no-movement blocks for each perturbation direction. A short self-timed break was allowed between each block. In order to encourage participants to produce their speech movements quickly, they were given points based on their response latency on movement trials. Ten points were given for responses with latencies up to 500 ms, falling off by 1 point every 40 ms thereafter. To encourage participants to inhibit overt speech production on no-movement trials, −25 points were given when a spoken response was detected. On all trials, the onset of speech (and response latency) was determined as the point where the amplitude of the microphone signal crossed above a predetermined, low amplitude threshold. Prior to the experiment, participants were trained in this paradigm: after a verbal description of the task, participants completed a single block of trials (21 total trials). This training was repeated as needed until participants showed the ability to respond appropriately on the no-movement trials (training repeated if participants produced any vocalization on more than 2/4 no-movement trials). A second experiment (Experiment 2) followed the same design, with “observation” trials used instead of “no-movement” trials (Fig 4A). In this experiment, either green or pink visual cues were presented before stimulus words. The green picture of someone speaking indicates “speak,” and participants were instructed to read the words aloud, while the pink picture of someone with their mouth closed indicates “don’t speak,” and participants were instructed to remain still. Since these cues were presented before each stimulus word, observation trials would not elicit movement planning, and participants instead passively listened to a recording of their own speech from the previous trial with the formant perturbation applied, as in “no-movement” trials in Experiment 1. Data quantification Formant data were tracked using wave_viewer [69], which provides a MATLAB GUI interface for formant tracking using Praat [70]. LPC order and pre-emphasis values were set individually for each participant. Vowels were initially automatically identified by locating the samples which were above a participant-specific amplitude level. Subsequently, all trials were hand-checked for errors. Errors in formant tracking, identified by discontinuities in the formant tracks or values outside a vowel’s typical range, were corrected by adjusting the pre-emphasis value or LPC order. Errors in the location of vowel onset and offset were corrected by hand-marking these times using landmarks in the audio waveform and spectrogram. For each trial, F1 was averaged from 50 to 125 ms after vowel onset (Fig 1E) to avoid (1) coarticulatory influences on vowel formants from the initial consonant and (2) potential changes in vowel formants due to online feedback corrections, which begin roughly 150 ms after vowel onset [17, 71]. Recent work has shown that this window is the most likely to accurately capture learning in this single-exposure paradigm [29]. Adaptation was quantified in three ways. Single-trial adaptation was measured as the change in F1 (in mels) from the first to the third trial of a triplet (i.e., after exposure to the auditory perturbation). We additionally calculated single-trial aftereffects—the retention of learning—by measuring the change in F1 (in mels) from the first trial in a triplet to the trial immediately following the third trial of that triplet. Finally, we computed an “adaptation index,” which was quantified by simply combining the above two metrics for each triplet. In all cases, the critical trials were unperturbed movement trials, either the first trial of the following triplet or a distractor trial. Because of the rather high variance in F1 production for vowels (standard deviation of ~15–20 Hz [72]) relative to the maximum expected effect size given previous work (~2 Hz, [29]), the adaptation index was our primary dependent variable of interest, designed to maximize our statistical power and potential to detect a small effect relative to the expected variance in production. A small number of trials were excluded due to errors in production (i.e., the participant said the wrong word), disfluencies, or unresolvable errors in formant tracking (discontinuities in formant tracks or cases where estimated formants clearly deviated from formants visible on the spectrogram): Experiment 1: median of 0.03% (0%–3.6% across participants), Experiment 2: median of 0.07% (0%–6% across participants). Additionally, all trials were manually examined by trained research assistants both visually (through the waveform and spectrogram) and auditorily to identify any sign of overt speech production, including stop burst release, vocalization, and/or frication. No-movement triplets or observation triplets where participants produced any detectable vocalization on the middle, perturbed, trial were excluded in order to isolate learning effects when no speech movement was produced (Experiment 1: 4–89 triplets excluded across participants, median 32, Experiment 2: 0–115 triplets excluded, median 3). In total, these exclusions resulted in 64–72 independent measurements of movement triplets (median 71) and 52–140 independent measurements of no-movement triplets for each participant (median 111) in Experiment 1 and 63–72 independent measurements of movement triplets (median 70) and 127–140 independent measurements of observation triplets (median 137) in Experiment 2. Statistical analysis Following our previous study in reaching [28], linear mixed effects models were run using the lme4 package in R [73]. For single-trial adaptation, retention, and the adaptation index models, model predictors included perturbation direction (upward or downward), triplet type (movement or no-movement), and the interaction between these two factors. All trials were included in the model, with Participant was included as a random effect (intercept) to control for multiple observations from each individual participant. Statistical significance was assessed using the lmerTest package [74]. Post-hoc comparisons were conducted with estimated marginal means using the Satterthwaite method for approximating the degrees of freedom [75]. Effect sizes are reported as partial R2 values from the linear mixed effects models and Cohen’s d for pairwise comparisons. Summary statistics report means and standard errors. References - 1. Wolpert DM, Kawato M. Multiple paired forward and inverse models for motor control. Neural Netw. 1998;11(7–8):1317–29. pmid:12662752 - 2. Wolpert DM, Flanagan JR. Computations underlying sensorimotor learning. Curr Opin Neurobiol. 2016;37:7–11. pmid:26719992 - 3. Krakauer JW, Hadjiosif AM, Xu J, Wong AL, Haith AM. Motor Learning. Compr Physiol. 2019;9(2):613–63. pmid:30873583 - 4. Scott SH. Optimal feedback control and the neural basis of volitional motor control. Nat Rev Neurosci. 2004;5(7):532–46. pmid:15208695 - 5. Shadmehr R, Krakauer JW. A computational neuroanatomy for motor control. Exp Brain Res. 2008;185(3):359–81. pmid:18251019 - 6. Houde JF, Nagarajan SS. Speech production as state feedback control. Front Hum Neurosci. 2011;5:82. pmid:22046152 - 7. Parrell B, Ramanarayanan V, Nagarajan S, Houde J. The FACTS model of speech motor control: fusing state estimation and task-based control. PLoS Comput Biol. 2019;15(9):e1007321. pmid:31479444 - 8. Day KA, Roemmich RT, Taylor JA, Bastian AJ. Visuomotor learning generalizes around the intended movement. eNeuro. 2016;3. - 9. Sheahan HR, Franklin DW, Wolpert DM. Motor planning, not execution, separates motor memories. Neuron. 2016;92:773–9. - 10. Sheahan HR, Ingram JN, Žalalytė GM, Wolpert DM. Imagery of movements immediately following performance allows learning of motor skills that interfere. Sci Rep. 2018;8(1):14330. pmid:30254381 - 11. McDougle SD, Bond KM, Taylor JA. Implications of plan-based generalization in sensorimotor adaptation. J Neurophysiol. 2017. - 12. Vyas S, O’Shea DJ, Ryu SI, Shenoy KV. Causal role of motor preparation during error-driven learning. Neuron. 2020;106(2):329-339.e4. pmid:32053768 - 13. Tourville JA, Guenther FH. The DIVA model: a neural theory of speech acquisition and production. Lang Cogn Process. 2011;26(7):952–81. pmid:23667281 - 14. Guenther FH. Neural control of speech. Cambridge, MA: The MIT Press; 2016. - 15. Kim KS, Gaines JL, Parrell B, Ramanarayanan V, Nagarajan SS, Houde JF. Mechanisms of sensorimotor adaptation in a hierarchical state feedback control model of speech. PLoS Comput Biol. 2023;19(7):e1011244. pmid:37506120 - 16. Kim KS, Hinkley LB, Brent K, Gaines JL, Pongos AL, Gupta S, et al. Neurophysiological evidence of sensory prediction errors driving speech sensorimotor adaptation. J Neurosci. 2025;45(27):e2084242025. pmid:40425317 - 17. Parrell B, Agnew Z, Nagarajan S, Houde J, Ivry RB. Impaired feedforward control and enhanced feedback control of speech in patients with cerebellar degeneration. J Neurosci. 2017;37(38):9249–58. pmid:28842410 - 18. Parrell B, Kim HE, Breska A, Saxena A, Ivry R. Differential effects of cerebellar degeneration on feedforward versus feedback control across speech and reaching movements. J Neurosci. 2021;41(42):8779–89. pmid:34446570 - 19. Cowie R, Douglas-Cowie E, Kerr AG. A study of speech deterioration in post-lingually deafened adults. J Laryngol Otol. 1982;96(2):101–12. pmid:7057081 - 20. Lane H, Webster JW. Speech deterioration in postlingually deafened adults. J Acoust Soc Am. 1991;89(2):859–66. pmid:2016436 - 21. Svirsky MA, Lane H, Perkell JS, Wozniak J. Effects of short-term auditory deprivation on speech production in adult cochlear implant users. J Acoust Soc Am. 1992;92(3):1284–300. pmid:1401516 - 22. Perkell JS, Guenther FH, Lane H, Matthies ML, Perrier P, Vick J, et al. A theory of speech motor control and supporting data from speakers with normal hearing and with profound hearing loss. J Phon. 2000;28:233–72. - 23. Wolpert DM, Diedrichsen J, Flanagan JR. Principles of sensorimotor learning. Nat Rev Neurosci. 2011;12(12):739–51. pmid:22033537 - 24. Howard IS, Wolpert DM, Franklin DW. The effect of contextual cues on the encoding of motor memories. J Neurophysiol. 2013;109(10):2632–44. pmid:23446696 - 25. Heald JB, Ingram JN, Flanagan JR, Wolpert DM. Multiple motor memories are learned to control different points on a tool. Nat Hum Behav. 2018;2(4):300–11. pmid:29736420 - 26. Avraham G, Taylor JA, Breska A, Ivry RB, McDougle SD. Contextual effects in sensorimotor adaptation adhere to associative learning rules. Elife. 2022;11:e75801. pmid:36197002 - 27. Churchland MM, Shenoy KV. Preparatory activity and the expansive null-space. Nat Rev Neurosci. 2024;25(4):213–36. pmid:38443626 - 28. Kim OA, Forrence AD, McDougle SD. Motor learning without movement. Proc Natl Acad Sci U S A. 2022;119(30):e2204379119. pmid:35858450 - 29. Hantzsch L, Parrell B, Niziolek CA. A single exposure to altered auditory feedback causes observable sensorimotor adaptation in speech. Elife. 2022;11:e73694. pmid:35816163 - 30. Mattar AAG, Gribble PL. Motor learning by observing. Neuron. 2005;46(1):153–60. pmid:15820701 - 31. Pawlowsky C, Thénault F, Bernier PF. Implicit sensorimotor adaptation proceeds in absence of movement execution. eNeuro. 2023;10. - 32. Wolpert DM, Miall RC, Kawato M. Internal models in the cerebellum. Trends Cogn Sci. 1998;2(9):338–47. pmid:21227230 - 33. Smith MA, Ghazizadeh A, Shadmehr R. Interacting adaptive processes with different timescales underlie short-term motor learning. PLoS Biol. 2006;4(6):e179. pmid:16700627 - 34. Shadmehr R, Smith MA, Krakauer JW. Error correction, sensory prediction, and adaptation in motor control. Annu Rev Neurosci. 2010;33:89–108. pmid:20367317 - 35. Hadjiosif AM, Krakauer JW, Haith AM. Did we get sensorimotor adaptation wrong? Implicit adaptation as direct policy updating rather than forward-model-based learning. J Neurosci. 2021;41(12):2747–61. pmid:33558432 - 36. Kearney E, Nieto-Castañón A, Weerathunge HR, Falsini R, Daliri A, Abur D, et al. A simple 3-parameter model for examining adaptation in speech and voice production. Front Psychol. 2020;10. - 37. Kawato M, Furukawa K, Suzuki R. A hierarchical neural-network model for control and learning of voluntary movement. Biol Cybern. 1987;57(3):169–85. pmid:3676355 - 38. Albert ST, Shadmehr R. The neural feedback response to error as a teaching signal for the motor learning system. J Neurosci. 2016;36(17):4832–45. pmid:27122039 - 39. Rochet-Capellan A, Richer L, Ostry DJ. Nonhomogeneous transfer reveals specificity in speech motor learning. J Neurophysiol. 2012;107(6):1711–7. pmid:22190628 - 40. Caudrelier T, Schwartz JL, Perrier P, Gerber S, Rochet-Capellan A. Transfer of learning: what does it tell us about speech production units?. J Speech Lang Hear Res. 2018;1. - 41. Gandolfo F, Mussa-Ivaldi FA, Bizzi E. Motor learning by field approximation. Proc Natl Acad Sci U S A. 1996;93(9):3843–6. pmid:8632977 - 42. Krakauer JW, Pine ZM, Ghilardi MF, Ghez C. Learning of visuomotor transformations for vectorial planning of reaching trajectories. J Neurosci. 2000;20(23):8916–24. pmid:11102502 - 43. Donchin O, Francis JT, Shadmehr R. Quantifying generalization from trial-by-trial behavior of adaptive systems that learn with basis functions: theory and experiments in human motor control. J Neurosci. 2003;23(27):9032–45. pmid:14534237 - 44. Kitazawa S, Kohno T, Uka T. Effects of delayed visual information on the rate and amount of prism adaptation in the human. J Neurosci. 1995;15(11):7644–52. pmid:7472515 - 45. Brudner SN, Kethidi N, Graeupner D, Ivry RB, Taylor JA. Delayed feedback during sensorimotor learning selectively disrupts adaptation but not strategy use. J Neurophysiol. 2016;115:1499–511. - 46. Schween R, Hegele M. Feedback delay attenuates implicit but facilitates explicit adjustments to a visuomotor rotation. Neurobiol Learn Mem. 2017;140:124–33. pmid:28257877 - 47. Zhou W, Fitzgerald J, Colucci-Chang K, Murthy KG, Joiner WM. The temporal stability of visuomotor adaptation generalization. J Neurophysiol. 2017;118(4):2435–47. pmid:28768744 - 48. McDougle SD, Taylor JA. Dissociable cognitive strategies for sensorimotor learning. Nat Commun. 2019;10:40. - 49. Max L, Maffett DG. Feedback delays eliminate auditory-motor learning in speech production. Neurosci Lett. 2015;591:25–9. pmid:25676810 - 50. Shiller DM, Mitsuya T, Max L. Exposure to auditory feedback delay while speaking induces perceptual habituation but does not mitigate the disruptive effect of delay on speech auditory-motor learning. Neuroscience. 2020;446:213–24. pmid:32738430 - 51. Tsay JS, Kim H, Haith AM, Ivry RB. Understanding implicit sensorimotor adaptation as a process of proprioceptive re-alignment. Elife. 2022;11:e76639. pmid:35969491 - 52. Ebrahimi S, Ostry DJ. The human somatosensory cortex contributes to the encoding of newly learned movements. Proc Natl Acad Sci U S A. 2024;121(6):e2316294121. pmid:38285945 - 53. Curio G, Neuloh G, Numminen J, Jousmäki V, Hari R. Speaking modifies voice-evoked activity in the human auditory cortex. Hum Brain Mapp. 2000;9(4):183–91. pmid:10770228 - 54. Houde JF, Nagarajan SS, Sekihara K, Merzenich MM. Modulation of the auditory cortex during speech: an MEG study. J Cogn Neurosci. 2002;14(8):1125–38. pmid:12495520 - 55. Ventura MI, Nagarajan SS, Houde JF. Speech target modulates speaking induced suppression in auditory cortex. BMC Neurosci. 2009;10:58. pmid:19523234 - 56. Flinker A, Chang EF, Kirsch HE, Barbaro NM, Crone NE, Knight RT. Single-trial speech suppression of auditory cortex activity in humans. J Neurosci. 2010;30(49):16643–50. pmid:21148003 - 57. Niziolek CA, Nagarajan SS, Houde JF. What does motor efference copy represent? Evidence from speech production. J Neurosci. 2013;33(41):16110–6. pmid:24107944 - 58. Beach SD, Tang D-L, Kiran S, Niziolek CA. Pars opercularis underlies efferent predictions and successful auditory feedback processing in speech: evidence from left-hemisphere stroke. Neurobiol Lang (Camb). 2024;5(2):454–83. pmid:38911464 - 59. Tang D-L, Parrell B, Beach SD, Niziolek CA. The brain’s sensitivity to sensory error can be modulated by altering perceived variability. J Neurosci. 2025;45(5):e0024242024. pmid:39622646 - 60. Eliades SJ, Wang X. Sensory-motor interaction in the primate auditory cortex during self-initiated vocalizations. J Neurophysiol. 2003;89(4):2194–207. pmid:12612021 - 61. Daliri A, Max L. Modulation of auditory responses to speech vs. nonspeech stimuli during speech movement planning. Front Hum Neurosci. 2016;10:234. pmid:27242494 - 62. Mitsuya T, MacDonald EN, Munhall KG, Purcell DW. Formant compensation for auditory feedback with English vowels. J Acoust Soc Am. 2015;138(1):413–24. pmid:26233040 - 63. Niziolek CA, Parrell B. Responses to auditory feedback manipulations in speech may be affected by previous exposure to auditory errors. J Speech Lang Hear Res. 2021;64(6S):2169–81. pmid:33705674 - 64. Zeng Y, Niziolek CA, Parrell B. Simultaneous acquisition of multiple auditory-motor transformations reveals suprasyllabic motor planning in speech production. J Exp Psychol Gen. 2025;154(6):1681–98. pmid:40208724 - 65. Tilsen S, Goldstein L. Articulatory gestures are individually selected in production. J Phon. 2012;40(6):764–79. pmid:23002316 - 66. Cai S, Boucek M, Ghosh S, Guenther FH, Perkell J. A system for online dynamic perturbation of formant trajectories and results from perturbations of the Mandarin triphthong /iau/. In: Proceedings of the 8th International Seminar on Speech Production, Strasbourg, France, 2008. p. 65–8. - 67. Tourville JA, Cai S, Guenther F. Exploring auditory-motor interactions in normal and disordered speech. In: Proceedings of Meetings on Acoustics, 2013. 060180–060180. - 68. Kim KS, Wang H, Max L. It’s about time: minimizing hardware and software latencies in speech research with real-time auditory feedback. J Speech Lang Hear Res. 2020;63(8):2522–34. pmid:32640180 - 69. Niziolek CA, Houde J. Wave_Viewer: First Release. 2015 [cited 2020 Jan 30]. Available from: https://zenodo.org/record/13839 - 70. Boersma P, Weenink D. Praat: doing phonetics by computer. 2019. Available from: http://www.praat.org/ - 71. Cai S, Beal DS, Ghosh SS, Tiede MK, Guenther FH, Perkell JS. Weak responses to auditory feedback perturbation during articulation in persons who stutter: evidence for abnormal auditory-motor transformation. PLoS One. 2012;7(7):e41830. pmid:22911857 - 72. Whalen DH, Chen WR. Variability and central tendencies in speech production. Front Commun. 2019;4:49. - 73. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;67. - 74. Kuznetsova A, Brockhoff PB, Christensen RHB. lmerTest package: tests in linear mixed effects models. J Stat Softw. 2017;82. - 75. Lenth R, Singmann H, Love J, Buerkner P, Herve M. Emmeans: estimated marginal means, aka least-squares means. 2023. Available from: https://CRAN.R-project.org/package=emmeans