Speech and language therapy: A treatment case series of 20 patients with functional speech disorder

2.1 Chart review

We reviewed the medical records of 24 consecutive individuals referred for the first time to outpatient SLT from an outpatient FND clinic at the Massachusetts General Hospital beginning from October 2014 through the end of February 2021. All patients were diagnosed with clinically established FND-speech using rule-in features (American Psychiatric Association, 2013; Baker et al., 2021) at their FND clinic appointment as assessed by a neurologist or neuropsychiatrist, and this diagnosis was confirmed at the SLT evaluation (conducted for all patients by co-author J.F., a certified and licensed speech and language pathologist with 7 years of experience treating patients with FND-speech). Four of the 24 patients were excluded, as SLT treatment sessions were never initiated (3 lost to follow-up, 1 asymptomatic at evaluation). The remaining sample (n = 20), all fluent in English (2 of 20 individuals identified Spanish as their first language and selected English as their preferred language for healthcare), had ended treatment by September 2021. Given the focus on FND-speech, isolated functional voice disorders (FND-voice) and/or those referred to SLT for cognitive symptoms without a concurrent FND-speech were excluded (Baker, 2016).

Demographic characteristics and baseline clinical variables were extracted from FND clinic notes prior to SLT intake and from the evaluation note of the single speech and language pathologist. Variables included FND-specific clinical features (including FND-speech phenotype and the presence of other functional neurological symptoms), medical and neuropsychiatric characteristics (e.g., concurrent pain, fatigue, insomnia, lifetime depression, anxiety), and any other FND-specific concurrent treatment (e.g., physical therapy, psychotherapy).

2.2 SLT assessment

All patients participated in a full motor speech and language examination with perceptual analysis and description by the speech and language pathologist (co-author J.F.) as consistent with broad clinical practice recommendations for all motor speech disorders (American Speech-Language-Hearing Association, 2022). The initial assessment ranged from 60 to 120 minutes in length. Examination components included an in-depth case history obtained via chart review and interview, an oral mechanism examination, and a variety of speech production tasks including reading, oral picture description, and free conversation. In addition to clinical observations made during the session, speech samples were recorded and qualitatively analyzed. Perceptual analyses of speech samples focused on patient-specific metrics (e.g., voice quality, fluency, articulatory precision), as well as positive rule-in features of FND-speech including internal task-to-task inconsistency and distractibility (Baker et al., 2021; Utianski & Duffy, 2022). Initial trials of speech modification strategies to assess stimulability for change were typically performed at the assessment, albeit with some variability due to time constraints.

Importantly, the assessment and treatment course for this cohort partially overlapped with the COVID-19 pandemic. As such, virtual (video telehealth) care was at times offered in place of in-person treatment due to hospital policy and/or patient preferences in addition to the clinician’s determination of appropriateness. Virtual care sessions did not differ inherently from in-person sessions in terms of targets, scope, or frequency.

2.3 Definition of FND-speech impairments observed on SLT examination

Each patient’s speech difficulties on SLT examination were described in the clinician’s initial assessment based on the motor speech subsystems affected. Traditionally, the five speech subsystems include respiration (speech breathing), phonation (voice quality), articulation (formation of specific speech sounds), resonance (shaping of airflow by oral/nasal cavities), and prosody (stress, rhythm, and intonation of speech) (Duffy, 2019). Two additional characteristics, speech fluency and agrammatism, were also recorded due to their prominence in this population. For the purposes of data collection, fluency was defined as disruptions to the continuity and consistency of speech rate, and agrammatism as disruptions in the grammar and/or syntax of a patient’s verbal language output. When applicable, functional motor symptoms affecting speech (e.g., dis-coordination of orofacial structures) were also noted.

2.4 Definition of general FND-speech phenotypes

Three FND-speech phenotypes – functional stutter, prosodic disturbance/foreign accent, and articulatory disturbance/infantile speech – have been described in the literature as detailed earlier (Baker et al., 2021; Duffy, 2019; McWhirter et al., 2019). Each patient in this cohort exhibited hallmark characteristics of at least one of these three categorical descriptors, and corresponding phenotypes were recorded based on the speech and language pathologist’s clinical descriptions of patient presentation.

2.5 SLT treatment

Treatment sessions, whether conducted in the clinic or virtually, ranged from 45 to 60 minutes in length and were also conducted by the same speech and language pathologist (co-author J.F.). Therapeutic interventions paralleled the principles found in the SLT consensus recommendations for FND-speech (Baker et al., 2021), beginning with patient education and counseling followed by initiation of symptomatic treatment. The initial treatment session was often dedicated to deepening the patient’s understanding of their atypical speech production and the underlying mechanisms at play (e.g., excess effort, musculoskeletal tension). Once shared language and a common understanding of therapy goals were established, the patient participated in treatment tailored to modifying their atypical speech characteristics while concurrently addressing unhelpful thoughts and behaviors contributing to communication disruptions. Given the importance of this psychotherapeutically-informed counseling alongside the motor speech focused symptomatic interventions, the precise targets of each SLT session necessarily varied from patient to patient. The recommended frequency of sessions typically decreased over the course of treatment as commensurate with patient improvement, but frequency of sessions also varied based on other relevant clinical factors and patient availability.

Symptomatic treatment followed a stepwise progression. Initially, both general and symptom-specific strategies were used to facilitate improved speech output. Broadly used strategies included attempts to facilitate a return to “natural” speech patterns through automatic and reflexive sounds, while symptom-specific interventions were tailored to target the specific breakdowns in one or more speech subsystems (see Table 1). Also, see Freeburn, 2022; Utianski & Duffy, 2022 for a case-based illustrations of the treatment of FND-speech. Following successful facilitation of a more typical speech pattern, the treatment targeted application of these learned strategies to a hierarchy of progressively more complex and lengthier speech tasks (e.g., from repetition to reading to structured conversation). The patient’s demonstrated accuracy and self-monitoring of speech production at each respective level was used by the clinician to gauge readiness to advance within this hierarchy.

Once patients achieved the goal of producing accurate speech during unstructured conversation, further goals targeting carryover to daily communication environments were established and targeted within the clinic (e.g., applying skills to novel situations such as making a phone call from the SLT office), as well as outside of the clinic. The final (discharge) treatment session covered strategies for self-cuing in case of future symptom relapse.

For individuals with FND-speech and concurrent cognitive complaints, SLT treatment targeting metacognitive education and implementation of strategies to support cognitive communication functioning was also introduced during the therapy course using approaches detailed elsewhere (Freeburn, 2022).

Information on SLT outcomes was extracted from clinician notes on the last available SLT treatment session attended by the patient. The primary outcome measure was clinician-determined improvement based on a combination of perceptual measures and patient progress toward treatment goals. “Asymptomatic” patients were noted as being completely or essentially back to their baseline speech pattern as gauged by clinician impression. Patients rated as “improved” had chart-documented evidence of significant and applicable improvement with implementation of speech strategies during clinical tasks or general improvement in FND-speech symptoms over the course of treatment. When patients reached either substantial improvement or a notable plateau in progress, discharge from SLT was recommended. Henceforth, the term “at discharge” will be used to describe the outcomes detailed in either the SLT clinician’s discharge note or the patient’s last known follow-up visit, and “improved” patients will refer to those who improved but were not asymptomatic at discharge.

Patient outcomes extracted from SLT notes also included the type of treatments received and targets achieved by each patient at discharge. For example, an outcome at the “conversation level” indicates that the patient consistently applied learned speech strategies to produce accurate speech during >90% of utterances during structured clinical conversational tasks.

4.1 Treatment impact of non-motor symptoms and other psychosocial features

A notable observation is that a subset of patients in the asymptomatic and improved categories were “on or applying for disability” and/or endorsed other bodily symptoms such as pain, fatigue, and cognitive complaints. These factors in other FND populations have been associated with poor clinical outcomes and reduced quality of life (Gelauff & Stone, 2016; Glass et al., 2018; Maggio et al., 2020; Věchetová et al., 2018). For example, a concurrent diagnosis of somatization disorder (indicative of prominent bodily complaints including pain) correlated with poor prognosis in a 14-year prospective study of clinical outcomes in patients with functional weakness (Gelauff et al., 2019). In patients with FND-seiz, concurrent pain syndromes negatively correlated with quality of life (Myers et al., 2012); the presence of cognitive complaints has also been linked to decreased quality of life in patients with FND-movt (Věchetová et al., 2018). Given the lack of prospective controlled clinical trial research in FND-speech to date, our case series findings suggest that the inclusion of a broad spectrum of FND-speech cases would likely be feasible for initial SLT trials. Similarly, being “on or applying for disability” does not appear to preclude the appreciation of positive communication-related clinical outcomes in this population.

4.2 Virtual (Video Telehealth) SLT for FND-Speech

Given that 8 of 15 patients rated as improved at discharge received care virtually, this case series also lends support to a role for delivering SLT for FND-speech via video telehealth (Perez, 2021; Perez et al., 2020). Several pilot video telehealth studies have been performed in patients with FND. For example, a recently published study identified that use of 12 sessions of weekly physical therapy delivered via telemedicine following completion of a 5-day intensive rehabilitation program showed favorable clinical outcomes compared to self-management (Gandolfi et al., 2022). Other articles have indicated similar efficacy of virtual treatment (Demartini et al., 2020; Perez, 2021). Given the reduced prevalence of FND-speech in comparison to other subtypes of FND (e.g., FND-seiz, FND-movt) (Aybek & Perez, 2022), it is likely that leveraging video telehealth treatment delivery approaches will aid both participant recruitment for FND-speech focused clinical trials and the availability/dissemination of consensus treatment recommendations in the field (Baker et al., 2021).

4.3 Other considerations

From an FND-speech phenotypic characterization perspective, this case series raises some interesting questions. Most notably, across the 3 general phenotypes (i.e., functional stutter, prosodic disturbance/foreign accent, and articulatory impairments/infantile speech), 50% of individuals in this cohort had mixed speech phenotypes. Furthermore, 18 of 20 individuals with FND-speech had a concurrent functional motor symptom (across the spectrum of hyperkinetic and hypokinetic FND-movt that includes functional limb weakness); 4 of 20 individuals were also diagnosed with FND-seiz. The frequency of mixed FND presentations that includes functional speech symptoms as part of the overall symptom complex further supports the need for a transdiagnostic approach to the characterization and treatment of patients with FND that cuts across the FND-speech, FND-movt, and FND-seiz literatures (Lidstone et al., 2022; Matin et al., 2017; Tinazzi et al., 2020). Interesting differences to further investigate include that the average age of patients in this cohort was 51.6 years old, with a mean speech-related illness duration of 3 years. If such findings are replicated in large, prospectively collected cohorts, this would suggest that FND-speech onset occurs over a decade later than typical onset of functional dystonia and FND-seiz (Lidstone et al., 2022). Thus, while frequent mixed symptom presentations suggest convergent neural mechanisms and etiological considerations, age of onset differences lend support to some potential neurobiological differences that also warrant more research. Lastly, from an SLT perspective, it is notable that there is not a simple one-to-one correspondence between general FND-speech phenotypes as described in the literature and speech impairments detected on SLT examination. For example, patients exhibiting hallmark perceptual characteristics of “infantile speech” presented with speech impairments across multiple subsystems not limited to articulation. This finding highlights the need for broader integration of FND-speech perceptual phenotypes and speech impairments, perhaps resulting in further refinement of phenotype labels.

4.4 Limitations and future directions

There are several limitations to this present effort. This is a small, retrospectively collected patient sample that used only clinically available medical record data. While all patients in this cohort had speech difficulties impairing social and/or occupational functioning, we did not have a more specific clinician-rated or self-report measure of the severity of their functional speech symptoms (nor of their overall FND symptom complex more broadly). Future research should include prospective data collection with standardized baseline and post-treatment speech and language assessments (including long-term follow-up data), with a range of objective, clinician-rated, and patient-reported symptom severity measures (T. R. Nicholson et al., 2020; Pick et al., 2020). Pre-treatment psychometric characterization of neuropsychiatric variables will also aid future research regarding potential factors predicting response to SLT interventions for FND-speech. Furthermore, it is important to note that most patients in this FND-speech cohort were concurrently receiving other physical rehabilitation and/or psychotherapy interventions. While a multidisciplinary (and interdisciplinary) approach to patient care is encouraged for many patients with FND (LaFaver & Ricciardi, 2022), such naturalistic interventions make it difficult to explicitly link clinical improvement to one specific treatment modality. Future research is also encouraged to include a broad data analysis of all concurrent treatment progression, as it is unclear in this sample if FND-speech improvement corresponded with overall FND symptom complex improvement.