Artificial Intelligence to Differentiate Pediatric Pseudopapilledema and True Papilledema on Fundus Photographs.

Purpose: To develop and test an artificial intelligence (AI) model to aid in differentiating pediatric pseudopapilledema from true papilledema on fundus photographs. Design: Multicenter retrospective study. Subjects: A total of 851 fundus photographs from 235 children (age < 18 years) with pseudopapilledema and true papilledema. Methods: Four pediatric neuro-ophthalmologists at 4 different institutions contributed fundus photographs of children with confirmed diagnoses of papilledema or pseudopapilledema. An AI model to classify fundus photographs as papilledema or pseudopapilledema was developed using a DenseNet backbone and a tribranch convolutional neural network. We performed 10-fold cross-validation and separately analyzed an external test set. The AI model's performance was compared with 2 masked human expert pediatric neuro-ophthalmologists, who performed the same classification task. Main Outcome Measures: Accuracy, sensitivity, and specificity of the AI model compared with human experts. Results: The area under receiver operating curve of the AI model was 0.77 for the cross-validation set and 0.81 for the external test set. The accuracy of the AI model was 70.0% for the cross-validation set and 73.9% for the external test set. The sensitivity of the AI model was 73.4% for the cross-validation set and 90.4% for the external test set. The AI model's accuracy was significantly higher than human experts on the cross validation set (P < 0.002), and the model's sensitivity was significantly higher on the external test set (P = 0.0002). The specificity of the AI model and human experts was similar (56.4%-67.3%). Moreover, the AI model was significantly more sensitive at detecting mild papilledema than human experts, whereas AI and humans performed similarly on photographs of moderate-to-severe papilledema. On review of the external test set, only 1 child (with nearly resolved pseudotumor cerebri) had both eyes with papilledema incorrectly classified as pseudopapilledema. Conclusions: When classifying fundus photographs of pediatric papilledema and pseudopapilledema, our AI model achieved > 90% sensitivity at detecting papilledema, superior to human experts. Due to the high sensitivity and low false negative rate, AI may be useful to triage children with suspected papilledema requiring work-up to evaluate for serious underlying neurologic conditions. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

A machine learning approach to improve implementation monitoring of family-based preventive interventions in primary care.

Background: Evidence-based parenting programs effectively prevent the onset and escalation of child and adolescent behavioral health problems. When programs have been taken to scale, declines in the quality of implementation diminish intervention effects. Gold-standard methods of implementation monitoring are cost-prohibitive and impractical in resource-scarce delivery systems. Technological developments using computational linguistics and machine learning offer an opportunity to assess fidelity in a low burden, timely, and comprehensive manner. Methods: In this study, we test two natural language processing (NLP) methods [i.e., Term Frequency-Inverse Document Frequency (TF-IDF) and Bidirectional Encoder Representations from Transformers (BERT)] to assess the delivery of the Family Check-Up 4 Health (FCU4Health) program in a type 2 hybrid effectiveness-implementation trial conducted in primary care settings that serve primarily Latino families. We trained and evaluated models using 116 English and 81 Spanish-language transcripts from the 113 families who initiated FCU4Health services. We evaluated the concurrent validity of the TF-IDF and BERT models using observer ratings of program sessions using the COACH measure of competent adherence. Following the Implementation Cascade model, we assessed predictive validity using multiple indicators of parent engagement, which have been demonstrated to predict improvements in parenting and child outcomes. Results: Both TF-IDF and BERT ratings were significantly associated with observer ratings and engagement outcomes. Using mean squared error, results demonstrated improvement over baseline for observer ratings from a range of 0.83-1.02 to 0.62-0.76, resulting in an average improvement of 24%. Similarly, results demonstrated improvement over baseline for parent engagement indicators from a range of 0.81-27.3 to 0.62-19.50, resulting in an approximate average improvement of 18%. Conclusions: These results demonstrate the potential for NLP methods to assess implementation in evidence-based parenting programs delivered at scale. Future directions are presented. Trial registration: NCT03013309 ClinicalTrials.gov.

Research has shown that evidence-based parenting programs effectively prevent the onset and escalation of child and adolescent behavioral health problems. However, if they are not implemented with fidelity, there is a potential that they will not produce the same effects. Gold-standard methods of implementation monitoring include observations of program sessions. This is expensive and difficult to implement in delivery settings with limited resources. Using data from a trial of the Family Check-Up 4 Health program in primary care settings that served Latino families, we investigated the potential to make use of a form of machine learning called natural language processing (NLP) to monitor program delivery. NLP-based ratings were significantly associated with independent observer ratings of fidelity and participant engagement outcomes. These results demonstrate the potential for NLP methods to monitor implementation in evidence-based parenting programs delivered at scale.

SocialBit: protocol for a prospective observational study to validate a wearable social sensor for stroke survivors with diverse neurological abilities.

INTRODUCTION: Social isolation has been found to be a significant risk factor for health outcomes, on par with traditional risk factors. This isolation is characterised by reduced social interactions, which can be detected acoustically. To accomplish this, we created a machine learning algorithm called SocialBit. SocialBit runs on a smartwatch and detects minutes of social interaction based on vocal features from ambient audio samples without natural language processing. METHODS AND ANALYSIS: In this study, we aim to validate the accuracy of SocialBit in stroke survivors with varying speech, cognitive and physical deficits. Training and testing on persons with diverse neurological abilities allows SocialBit to be a universally accessible social sensor. We are recruiting 200 patients and following them for up to 8 days during hospitalisation and rehabilitation, while they wear a SocialBit-equipped smartwatch and engage in naturalistic daily interactions. Human observers tally the interactions via a video livestream (ground truth) to analyse the performance of SocialBit against it. We also examine the association of social interaction time with stroke characteristics and outcomes. If successful, SocialBit would be the first social sensor available on commercial devices for persons with diverse abilities. ETHICS AND DISSEMINATION: This study has received ethical approval from the Institutional Review Board of Mass General Brigham (Protocol #2020P003739). The results of this study will be published in a peer-reviewed journal.

The Silent Treatment?: Changes in patient emotional expression after silence.

Psychotherapy can be an emotionally laden conversation, where both verbal and non-verbal interventions may impact the therapeutic process. Prior research has postulated mixed results in how clients emotionally react following a silence after the therapist is finished talking, potentially due to studying a limited range of silences with primarily qualitative and self-report methodologies. A quantitative exploration may illuminate new findings. Utilizing research and automatic data processing from the field of linguistics, we analysed the full range of silence lengths (0.2 to 24.01 seconds), and measures of emotional expression - vocally encoded arousal and emotional valence from the works spoken - of 84 audio recordings Motivational Interviewing sessions. We hypothesized that both the level and the variance of client emotional expression would change as a function of silence length, however, due to the mixed results in the literature the direction of emotional change was unclear. We conducted a multilevel linear regression to examine how the level of client emotional expression changed across silence length, and an ANOVA to examine the variability of client emotional expression across silence lengths. Results indicated in both analyses that as silence length increased, emotional expression largely remained the same. Broadly, we demonstrated a weak connection between silence length and emotional expression, indicating no persuasive evidence that silence leads to client emotional processing and expression.

Creating musical features using multi-faceted, multi-task encoders based on transformers.

Computational machine intelligence approaches have enabled a variety of music-centric technologies in support of creating, sharing and interacting with music content. A strong performance on specific downstream application tasks, such as music genre detection and music emotion recognition, is paramount to ensuring broad capabilities for computational music understanding and Music Information Retrieval. Traditional approaches have relied on supervised learning to train models to support these music-related tasks. However, such approaches require copious annotated data and still may only provide insight into one view of music-namely, that related to the specific task at hand. We present a new model for generating audio-musical features that support music understanding, leveraging self-supervision and cross-domain learning. After pre-training using masked reconstruction of musical input features using self-attention bidirectional transformers, output representations are fine-tuned using several downstream music understanding tasks. Results show that the features generated by our multi-faceted, multi-task, music transformer model, which we call M3BERT, tend to outperform other audio and music embeddings on several diverse music-related tasks, indicating the potential of self-supervised and semi-supervised learning approaches toward a more generalized and robust computational approach to modeling music. Our work can offer a starting point for many music-related modeling tasks, with potential applications in learning deep representations and enabling robust technology applications.

Multimodal neuroimaging data from a 5-week heart rate variability biofeedback randomized clinical trial.

We present data from the Heart Rate Variability and Emotion Regulation (HRV-ER) randomized clinical trial testing effects of HRV biofeedback. Younger (N = 121) and older (N = 72) participants completed baseline magnetic resonance imaging (MRI) including T1-weighted, resting and emotion regulation task functional MRI (fMRI), pulsed continuous arterial spin labeling (PCASL), and proton magnetic resonance spectroscopy (1H MRS). During fMRI scans, physiological measures (blood pressure, pulse, respiration, and end-tidal CO2) were continuously acquired. Participants were randomized to either increase heart rate oscillations or decrease heart rate oscillations during daily sessions. After 5 weeks of HRV biofeedback, they repeated the baseline measurements in addition to new measures (ultimatum game fMRI, training mimicking during blood oxygen level dependent (BOLD) and PCASL fMRI). Participants also wore a wristband sensor to estimate sleep time. Psychological assessment comprised three cognitive tests and ten questionnaires related to emotional well-being. A subset (N = 104) provided plasma samples pre- and post-intervention that were assayed for amyloid and tau. Data is publicly available via the OpenNeuro data sharing platform.

A Machine-Learning Algorithm for the Automated Perceptual Evaluation of Dysphonia Severity.

OBJECTIVES: Auditory-perceptual assessments are the gold standard for assessing voice quality. This project aims to develop a machine-learning model for measuring perceptual dysphonia severity of audio samples consistent with assessments by expert raters. METHODS: The Perceptual Voice Qualities Database samples were used, including sustained vowel and Consensus Auditory-Perceptual Evaluation of Voice sentences, which were previously expertly rated on a 0-100 scale. The OpenSMILE (audEERING GmbH, Gilching, Germany) toolkit was used to extract acoustic (Mel-Frequency Cepstral Coefficient-based, n = 1428) and prosodic (n = 152) features, pitch onsets, and recording duration. We utilized a support vector machine and these features (n = 1582) for automated assessment of dysphonia severity. Recordings were separated into vowels (V) and sentences (S) and features were extracted separately from each. Final voice quality predictions were made by combining the features extracted from the individual components with the whole audio (WA) sample (three file sets: S, V, WA). RESULTS: This algorithm has a high correlation (r = 0.847) with estimates of expert raters. The root mean square error was 13.36. Increasing signal complexity resulted in better estimation of dysphonia, whereby combining the features outperformed WA, S, and V sets individually. CONCLUSION: A novel machine-learning algorithm was able to perform perceptual estimates of dysphonia severity using standardized audio samples on a 100-point scale. This was highly correlated to expert raters. This suggests that ML algorithms could offer an objective method for evaluating voice samples for dysphonia severity.

Wearable and Mobile Technologies for the Evaluation and Treatment of Obsessive-Compulsive Disorder: Scoping Review.

BACKGROUND: Smartphones and wearable biosensors can continuously and passively measure aspects of behavior and physiology while also collecting data that require user input. These devices can potentially be used to monitor symptom burden; estimate diagnosis and risk for relapse; predict treatment response; and deliver digital interventions in patients with obsessive-compulsive disorder (OCD), a prevalent and disabling psychiatric condition that often follows a chronic and fluctuating course and may uniquely benefit from these technologies. OBJECTIVE: Given the speed at which mobile and wearable technologies are being developed and implemented in clinical settings, a continual reappraisal of this field is needed. In this scoping review, we map the literature on the use of wearable devices and smartphone-based devices or apps in the assessment, monitoring, or treatment of OCD. METHODS: In July 2022 and April 2023, we conducted an initial search and an updated search, respectively, of multiple databases, including PubMed, Embase, APA PsycINFO, and Web of Science, with no restriction on publication period, using the following search strategy: ("OCD" OR "obsessive" OR "obsessive-compulsive") AND ("smartphone" OR "phone" OR "wearable" OR "sensing" OR "biofeedback" OR "neurofeedback" OR "neuro feedback" OR "digital" OR "phenotyping" OR "mobile" OR "heart rate variability" OR "actigraphy" OR "actimetry" OR "biosignals" OR "biomarker" OR "signals" OR "mobile health"). RESULTS: We analyzed 2748 articles, reviewed the full text of 77 articles, and extracted data from the 25 articles included in this review. We divided our review into the following three parts: studies without digital or mobile intervention and with passive data collection, studies without digital or mobile intervention and with active or mixed data collection, and studies with a digital or mobile intervention. CONCLUSIONS: Use of mobile and wearable technologies for OCD has developed primarily in the past 15 years, with an increasing pace of related publications. Passive measures from actigraphy generally match subjective reports. Ecological momentary assessment is well tolerated for the naturalistic assessment of symptoms, may capture novel OCD symptoms, and may also document lower symptom burden than retrospective recall. Digital or mobile treatments are diverse; however, they generally provide some improvement in OCD symptom burden. Finally, ongoing work is needed for a safe and trusted uptake of technology by patients and providers.

Modeling inter-individual differences in ambulatory-based multimodal signals via metric learning: a case study of personalized well-being estimation of healthcare workers.

Introduction: Intelligent ambulatory tracking can assist in the automatic detection of psychological and emotional states relevant to the mental health changes of professionals with high-stakes job responsibilities, such as healthcare workers. However, well-known differences in the variability of ambulatory data across individuals challenge many existing automated approaches seeking to learn a generalizable means of well-being estimation. This paper proposes a novel metric learning technique that improves the accuracy and generalizability of automated well-being estimation by reducing inter-individual variability while preserving the variability pertaining to the behavioral construct. Methods: The metric learning technique implemented in this paper entails learning a transformed multimodal feature space from pairwise similarity information between (dis)similar samples per participant via a Siamese neural network. Improved accuracy via personalization is further achieved by considering the trait characteristics of each individual as additional input to the metric learning models, as well as individual trait base cluster criteria to group participants followed by training a metric learning model for each group. Results: The outcomes of the proposed models demonstrate significant improvement over the other inter-individual variability reduction and deep neural baseline methods for stress, anxiety, positive affect, and negative affect. Discussion: This study lays the foundation for accurate estimation of psychological and emotional states in realistic and ambulatory environments leading to early diagnosis of mental health changes and enabling just-in-time adaptive interventions.

Mel frequency spectral domain defenses against adversarial attacks on speech recognition systems.

Automatic speech recognition (ASR) systems are vulnerable to adversarial attacks due to their reliance on machine learning models. Many of the defenses explored for defending ASR systems simply adapt defense approaches developed for the image domain. This paper explores speech-specific defenses in the feature domain and introduces a defense method called mel domain noise flooding (MDNF). MDNF injects additive noise to the mel spectrogram speech representation prior to re-synthesizing the audio signal input to ASR. The defense is evaluated against strong white-box threat models and shows competitive robustness.

It's not what you said, it's how you said it: An analysis of therapist vocal features during psychotherapy.

Psychotherapy is a conversation, whereby, at its foundation, many interventions are derived from the therapist talking. Research suggests that the voice can convey a variety of emotional and social information, and individuals may change their voice based on the context and content of the conversation (e.g., talking to a baby or delivering difficult news to patients with cancer). As such, therapists may adjust aspects of their voice throughout a therapy session depending on if they are beginning a therapy session and checking in with a client, conducting more therapeutic "work," or ending the session. In this study, we modeled three vocal features-pitch, energy, and rate-with linear and quadratic multilevel models to understand how therapists' vocal features change throughout a therapy session. We hypothesized that all three vocal features would be best fit with a quadratic function - starting high and more congruent with a conversational voice, decreasing during the middle portions of therapy where more therapeutic interventions were being administered, and increasing again at the end of the session. Results indicated a quadratic model for all three vocal features was superior in fitting the data, as compared to a linear model, suggesting that therapists begin and end therapy using a different style of voice than in the middle of a session.

Relationship satisfaction, feelings of closeness and annoyance, and linkage in electrodermal activity.

Physiological linkage refers to moment-to-moment, time-linked coordination in physiological responses among people in close relationships. Although people in romantic relationships have been shown to evidence linkage in their physiological responses over time, it is still unclear how patterns of covariation relate to in-the-moment, as well as general levels of, relationship functioning. In the present study with data collected between 2014 and 2017, we capture linkage in electrodermal activity (EDA) in a diverse sample of young-adult couples, generally representative and generalizable to the Los Angeles community from which we sampled. We test how naturally occurring, shifting feelings of closeness with and annoyance toward one's partner relate to concurrent changes in levels of physiological linkage over the course of 1 day. Additionally, we examine how linkage relates to overall relationship satisfaction. Results showed that couples evidenced significant covariation in their levels of physiological arousal in daily life. Further, physiological linkage increased during hours that participants felt close to their romantic partners but not during hours that participants felt annoyed with their partners. Finally, those participants with overall higher levels of relationship satisfaction showed lower levels of linkage over the day of data collection. These findings highlight how individuals respond in sync with their romantic partners and how this process ebbs and flows in conjunction with the shifting emotional tone of their relationships. The discussion focuses on how linkage might enhance closeness or, alternatively, contribute to conflict escalation and the potential of linkage processes to promote positive interpersonal relationships. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

You never know what you are going to get: Large-scale assessment of therapists' supportive counseling skill use.

Supportive counseling skills like empathy and active listening are critical ingredients of all psychotherapies, but most research relies on client or therapist reports of the treatment process. This study utilized machine-learning models trained to evaluate counseling skills to evaluate supportive skill use in 3,917 session recordings. We analyzed overall skill use and variation in practice patterns using a series of mixed effects models. On average, therapists scored moderately high on observer-rated empathy (i.e., 3.8 out of 5), 3.3% of the therapists' utterances in a session were open questions, and 12.9% of their utterances were reflections. However, there were substantial differences in skill use across therapists as well as across clients within-therapist caseloads. These findings highlight the substantial variability in the process of counseling that clients may experience when they access psychotherapy. We discuss findings in the context of both the need for therapists to be responsive and flexible with their clients, but also potential costs related to the lack of a more uniform experience of care. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Boys don't cry (or kiss or dance): A computational linguistic lens into gendered actions in film.

Contemporary media is full of images that reflect traditional gender notions and stereotypes, some of which may perpetuate harmful gender representations. In an effort to highlight the occurrence of these adverse portrayals, researchers have proposed machine-learning methods to identify stereotypes in the language patterns found in character dialogues. However, not all of the harmful stereotypes are communicated just through dialogue. As a complementary approach, we present a large-scale machine-learning framework that automatically identifies character's actions from scene descriptions found in movie scripts. For this work, we collected 1.2+ million scene descriptions from 912 movie scripts, with more than 50 thousand actions and 20 thousand movie characters. Our framework allow us to study systematic gender differences in movie portrayals at a scale. We show this through a series of statistical analyses that highlight differences in gender portrayals. Our findings provide further evidence to claims from prior media studies including: (i) male characters display higher agency than female characters; (ii) female actors are more frequently the subject of gaze, and (iii) male characters are less likely to display affection. We hope that these data resources and findings help raise awareness on portrayals of character actions that reflect harmful gender stereotypes, and demonstrate novel possibilities for computational approaches in media analysis.

Phone duration modeling for speaker age estimation in children.

Automatic inference of paralinguistic information from speech, such as age, is an important area of research with many technological applications. Speaker age estimation can help with age-appropriate curation of information content and personalized interactive experiences. However, automatic speaker age estimation in children is challenging due to the paucity of speech data representing the developmental spectrum, and the large signal variability including within a given age group. Most prior approaches in child speaker age estimation adopt methods directly drawn from research on adult speech. In this paper, we propose a novel technique that exploits temporal variability present in children's speech for estimation of children's age. We focus on phone durations as biomarker of children's age. Phone duration distributions are derived by forced-aligning children's speech with transcripts. Regression models are trained to predict speaker age among children studying in kindergarten up to grade 10. Experiments on two children's speech datasets are used to demonstrate the robustness and portability of proposed features over multiple domains of varying signal conditions. Phonemes contributing most to estimation of children speaker age are analyzed and presented. Experimental results suggest phone durations contain important development-related information of children. The proposed features are also suited for application under low data scenarios.

TILES-2019: A longitudinal physiologic and behavioral data set of medical residents in an intensive care unit.

The TILES-2019 data set consists of behavioral and physiological data gathered from 57 medical residents (i.e., trainees) working in an intensive care unit (ICU) in the United States. The data set allows for the exploration of longitudinal changes in well-being, teamwork, and job performance in a demanding environment, as residents worked in the ICU for three weeks. Residents wore a Fitbit, a Bluetooth-based proximity sensor, and an audio-feature recorder. They completed daily surveys and interviews at the beginning and end of their rotation. In addition, we collected data from environmental sensors (i.e., Internet-of-Things Bluetooth data hubs) and obtained hospital records (e.g., patient census) and residents' job evaluations. This data set may be may be of interest to researchers interested in workplace stress, group dynamics, social support, the physical and psychological effects of witnessing patient deaths, predicting survey data from sensors, and privacy-aware and privacy-preserving machine learning. Notably, a small subset of the data was collected during the first wave of the COVID-19 pandemic.

Enhancing the quality of cognitive behavioral therapy in community mental health through artificial intelligence generated fidelity feedback (Project AFFECT): a study protocol.

BACKGROUND: Each year, millions of Americans receive evidence-based psychotherapies (EBPs) like cognitive behavioral therapy (CBT) for the treatment of mental and behavioral health problems. Yet, at present, there is no scalable method for evaluating the quality of psychotherapy services, leaving EBP quality and effectiveness largely unmeasured and unknown. Project AFFECT will develop and evaluate an AI-based software system to automatically estimate CBT fidelity from a recording of a CBT session. Project AFFECT is an NIMH-funded research partnership between the Penn Collaborative for CBT and Implementation Science and Lyssn.io, Inc. ("Lyssn") a start-up developing AI-based technologies that are objective, scalable, and cost efficient, to support training, supervision, and quality assurance of EBPs. Lyssn provides HIPAA-compliant, cloud-based software for secure recording, sharing, and reviewing of therapy sessions, which includes AI-generated metrics for CBT. The proposed tool will build from and be integrated into this core platform. METHODS: Phase I will work from an existing software prototype to develop a LyssnCBT user interface geared to the needs of community mental health (CMH) agencies. Core activities include a user-centered design focus group and interviews with community mental health therapists, supervisors, and administrators to inform the design and development of LyssnCBT. LyssnCBT will be evaluated for usability and implementation readiness in a final stage of Phase I. Phase II will conduct a stepped-wedge, hybrid implementation-effectiveness randomized trial (N = 1,875 clients) to evaluate the effectiveness of LyssnCBT to improve therapist CBT skills and client outcomes and reduce client drop-out. Analyses will also examine the hypothesized mechanism of action underlying LyssnCBT. DISCUSSION: Successful execution will provide automated, scalable CBT fidelity feedback for the first time ever, supporting high-quality training, supervision, and quality assurance, and providing a core technology foundation that could support the quality delivery of a range of EBPs in the future. TRIAL REGISTRATION: ClinicalTrials.gov; NCT05340738 ; approved 4/21/2022.

Interpersonal synchrony across vocal and lexical modalities in interactions involving children with autism spectrum disorder.

Quantifying behavioral synchrony can inform clinical diagnosis, long-term monitoring, and individualised interventions in neuro-developmental disorders characterized by deficit in communication and social interaction, such as autism spectrum disorder. In this work, three different objective measures of interpersonal synchrony are evaluated across vocal and linguistic communication modalities. For vocal prosodic and spectral features, dynamic time warping distance and squared cosine distance of (feature-wise) complexity are used, and for lexical features, word mover's distance is applied to capture behavioral synchrony. It is shown that these interpersonal vocal and linguistic synchrony measures capture complementary information that helps in characterizing overall behavioral patterns.

Leveraging Social Networks for the Assessment and Management of Neurological Patients.

Social networks are the persons surrounding a patient who provide support, circulate information, and influence health behaviors. For patients seen by neurologists, social networks are one of the most proximate social determinants of health that are actually accessible to clinicians, compared with wider social forces such as structural inequalities. We can measure social networks and related phenomena of social connection using a growing set of scalable and quantitative tools increasing familiarity with social network effects and mechanisms. This scientific approach is built on decades of neurobiological and psychological research highlighting the impact of the social environment on physical and mental well-being, nervous system structure, and neuro-recovery. Here, we review the biology and psychology of social networks, assessment methods including novel social sensors, and the design of network interventions and social therapeutics.

Variation in compensatory strategies as a function of target constriction degree in post-glossectomy speech.

Individuals who have undergone treatment for oral cancer oftentimes exhibit compensatory behavior in consonant production. This pilot study investigates whether compensatory mechanisms utilized in the production of speech sounds with a given target constriction location vary systematically depending on target manner of articulation. The data reveal that compensatory strategies used to produce target alveolar segments vary systematically as a function of target manner of articulation in subtle yet meaningful ways. When target constriction degree at a particular constriction location cannot be preserved, individuals may leverage their ability to finely modulate constriction degree at multiple constriction locations along the vocal tract.