Updated systematic-narrative review on communication intervention in Rett Syndrome: 2010–2022

Abstract

Due to loss of spoken language and resulting complex communication needs, people with Rett syndrome are obvious candidates for communication intervention. To advance evidence-based practice and guide future research efforts, we identified and summarized 16 communication intervention studies published since a previous 2009 review on this topic. Studies were summarized in terms of (a) participants, (b) dependent variables related to communication, (c) intervention characteristics, (d) outcomes, and (e) certainty of evidence. Across the 16 studies, intervention was provided to a total of 100 participants from 3 to 47 years of age. Half of the studies used systematic instruction to teach aided AAC. Other interventions and associated technologies included music therapy, eye tracking technology, and transcranial stimulation. Positive outcomes (e.g., using AAC devices to make requests and/or initiate social-communication interactions) were reported in 13 of the studies. These 16 new studies provide additional guidance on how to enhance the communicative functioning of people with Rett syndrome. Future research directions are highlighted.

Keywords:

• Communication intervention
• Rett syndrome
• systematic-narrative review

Rett syndrome is an X-linked genetic disorder associated with severe developmental regression (Amir et al., 1999; Neul et al., 2010; Sandweiss et al., 2020). While it does rarely occur in males (Reichow et al., 2015), it is considerably more prevalent in females; approximately 1 case for every 10,000 female births (Fehr et al., 2011). At this prevalence rate, Rett syndrome appears to be the single most common cause of profound/multiple disability in girls (Smeets et al., 2012; Tarquinio & Percy, 2016).

In many cases the diagnosis can be confirmed due to mutations on the MECP2 gene (Amir et al., 1999). Regression usually begins early (between 6 and 18 months of age) and progresses rapidly, leading almost always to significant impairments by 4 years of age (Einspieler & Marschik, 2019; Foxe et al., 2016). The developmental course of Rett syndrome has been characterized in terms of four clinical stages (Fabio et al., 2020; Neul et al., 2010; Percy & Glaze, 2017). Generally, it is at Stage I—from about 6 to 18 months of age—when concerns such as poor muscle tone and delayed motor development might first be noticed. Next, from about 1 to 3 years of age (Stage II) the children often show developmental regression along with appearance of seizures (usually after age 2), acquired microcephaly, hyperventilation, scoliosis, loss of purposeful hand use, and emergence of stereotyped hand mannerisms. As the child ages and progresses to Stages III and Stage IV, most children will eventually present with significant communication and motor impairments. In addition, scoliosis has been estimated to occur in 85% of individuals by 16 years of age (Percy & Glaze, 2017).

In addition to these clinical stages, two main types of Rett syndrome have been identified: classic and atypical (Operto et al., 2019). With certain atypical variants there can be some preserved speech and greater motor/ambulation ability (Marschik et al., 2014; Skjeldal, 2020). In terms of evaluating interventions aimed at enhancing the functioning of people with Rett syndrome, it might be of some importance to ascertain the participants’ type and stage of Rett syndrome. Interventions effective for children in Stage III, for example, might not be useful at Stage II; however, there does not appear to be any research demonstrating intervention by type/stage interactions.

A number of studies have reported on the extent of spoken language abilities that remain as the disease progresses. Neul et al. (2014), for example, reported that the ability to communicate using phrases and single words was lost by nearly all the 638 females with both classic (n = 542) and atypical (n = 96) Rett syndrome in their longitudinal study. This finding is consistent with multiple other reports indicating that regressive (and extensive) loss of spoken language ability occurs for most children with both classic and atypical Rett syndrome (Bartolotta et al., 2011; Foxe et al., 2016; Hagberg, 2002; Marschik et al., 2014).

In light of these reports of significant spoken language impairment, it is not surprising that communication intervention, including augmentative and alternative communication (AAC) intervention, is a major treatment priority for people with Rett syndrome (Larriba-Quest et al., 2020). In short, people with Rett syndrome have complex communication needs and are likely to require extensive communication supports, including interventions aimed at teaching alternatives to speech (Light & McNaughton, 2011) and/or interventions aimed at expanding and strengthening the person’s existing communication skills (Julien et al., 2015).

Meeting the communication needs of people with Rett syndrome is likely to be challenging due to several unique characteristics associated with this condition. Along with communication and motor difficulties—which are challenging in their own right—Rett syndrome is also characterized by scoliosis, seizures, fluctuating levels of alertness, loss of purposeful hand use, and emergence of stereotyped mannerisms, such as hand mouthing, clapping, and wringing/washing-like hand movements (Ferreira & Teive, 2020; Neul et al., 2010). This latter characteristic will likely complicate and perhaps even contraindicate the use of certain alternative communication modes that require precise voluntary control of the fingers, hands, and/or arms, such as manual signing, picture exchange, and pointing to graphic symbols.

Given these unique challenges, an important question is whether there are in fact any empirically-validated (and successful) interventions for enhancing the communication functioning of people with Rett syndrome. A 2009 review by Sigafoos et al. offered a preliminary answer to this question. That review identified nine studies that collectively provided communication intervention to 31 children (from 2 to 17 years of age) with Rett syndrome. Three communication modes (i.e., speech, gestures, and selection of graphic symbols from a communication board or speech-generating device) were targeted for instruction across the nine studies. Participants were taught to use the targeted communication mode for a range of functions, primarily (a) requesting preferred objects/interaction, (b) naming/commenting, (c) imitating speech, and (d) receptive language tasks (e.g., responding to another’s request, answering questions, and pointing to named symbols). Intervention generally involved applying systematic instructional tactics (e.g., response prompting and reinforcement) or following a general therapeutic approach, such as providing music therapy or early intensive behavioral intervention. Improved communication outcomes (e.g., increased used of AAC symbols to make choices, request preferred objects, comment, name pictures, and respond to questions) were reported for 24 of the 31 children (84%); however, methodological limitations rendered the results inconclusive for all but one of the studies and that study had a sample size of only three children; thus, as of 2009, there were relatively few studies on providing communication intervention to people with Rett syndrome and most of these had methodological limitations.

The present paper updates the earlier (2009) review by examining studies published from 2010 to 2022. The general aim of this review was to document the extent and nature of current communication intervention literature for people with Rett syndrome. The specific aims were to (a) ascertain what, if any, new data are available on intervening to enhance the communication functioning of people with Rett syndrome; and (b) provide a narrative summary of the studies carried out in the area. Of particular interest was whether studies provided more conclusive evidence related to some of the promising practices identified in the 2009 review. Of additional interest was whether the ensuing 12 years had resulted in any significant advances in the literature, such as expansion of the literature to adults and to a wider range of communication modes and functions. An updated review of this type could facilitate practice and stimulate future research on communicative supports for people with Rett syndrome.

Method

The procedures adopted for this review were informed by the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA; Page et al., 2021).

Inclusion and exclusion criteria

To be included in this review, the study had to have evaluated the effects of an intervention that was aimed at teaching, increasing, and/or strengthening one or more areas of communication functioning. The intervention also had to have been applied to at least one person with Rett syndrome. Communication intervention was defined as implementing one or more therapeutic/teaching procedures for the purpose of trying to increase or improve the person’s communicative functioning, skills, or abilities. Intervention studies not specifically focused on teaching or increasing communication skills, but which included at least one dependent variable related to communicative functioning (e.g., scores on a standardized measure of receptive and expressive communication), were included.

A number of exclusion criteria were applied. Review papers and papers providing general intervention guidance, but which did not evaluate the effects of an intervention, were excluded. Studies that assessed or described the communication abilities, skills, and/or needs of people with Rett syndrome (e.g., assessing choice making abilities) without providing an intervention to develop or teach communication skills were also excluded. This exclusion extended to studies that made use of the child’s existing communication skills during an assessment. Studies that focused on the perceptions and/or experiences of parents/caregivers or professionals regarding the use of a particular communication intervention or method (e.g., parents’ perceptions and experiences with eye gaze technology) were excluded. Finally, studies that included participants both with and without Rett syndrome were excluded if it was not possible to isolate the results for the participants with Rett syndrome (e.g., if only means for the overall participant pool were reported).

Search and screening methods

Searches were undertaken in the following five electronic databases: Cumulative Index of Nursing and Allied Health Literatures (CINAHL), Education Resources Information Center (ERIC), Linguistics and Language Behavior Abstracts (LLBA), Medline, and PsycInfo. The search was limited to peer-reviewed journal articles published in English from 2010 onwards. The database search was conducted by the second author on May 11 2022.

Based on advice from a University of Newcastle librarian, the CINAHL, Medline, and PsycInfo databases were searched using the term “communication” and, separately, the term “Rett syndrome.” These search terms were entered into the Anywhere field. These two sets of terms were first searched separately and then combined with the limiting parameters applied. For the ERIC and LLBA databases, we conducted a single advanced search using the combined terms “communication” and “Rett syndrome.” These terms were also entered into the Anywhere field and the same limiting parameters used in the other databases were applied.

A total of 317 records were returned from the initial database search. Following the removal of duplicates (n = 132), the titles and abstracts of 185 records from the database search were independently screened by the first and second author. The purpose of this screening was to remove nonintervention studies, such as commentaries, review articles, and assessment-only studies. There was 98% agreement (four disagreements) on this initial screening. To resolve these four disagreements, the second author summarized each of these four studies and explained her rationale for including or excluding each of the four studies. The first author then responded by either agreeing with the second author or providing a rationale for any differing opinion. Through this process, the first and second author reached a consensus on the four initial disagreements. This resulted in 16 records being retained for full-text screening. Full-text screening was completed independently by the first and second authors. There was one disagreement (94% agreement) at this step. After reaching consensus on this disagreement, there was agreement between the first and second author that 13 studies met the inclusion criteria.

Subsequent hand searches were conducted of the reference lists for the 13 included studies. This occurred in June 2022. We also sent an e-mail to the lead or corresponding authors of the 13 already included studies. The e-mail informed them of our review and of their study or studies that had already been identified for inclusion. We asked if they had any additional papers on communication intervention for people with Rett syndrome for possible inclusion. These emails were sent on June 3rd and 4th, 2022 and 10 working days were allowed for responses. These subsequent search processes resulted in three additional studies being included; thus, a total of 16 studies were included in this review. Figure 1 provides a flowchart of the search and screening process.

Figure 1. Flowchart of process used to identify and screen studies for inclusion.

Data extraction

The 16 included studies were summarized in terms of (a) participants, (b) dependent variables related to communication, (c) intervention characteristics, and (d) main intervention outcomes. The first author extracted this information from each study and wrote a narrative summary of each study. The second author independently reviewed these summaries and any discrepancies were resolved by referring back to the original articles. We also contacted three author groups via email to clarify (a) procedural integrity checks (Byiers et al., 2014; Wandin et al., 2021) and (b) the version of the Vineland assessment that was used (Chou et al., 2019).

Certainty of evidence

The certainty of evidence for each study was rated as either conclusive, suggestive, or inconclusive based on the design and adequacy of reliability and procedural integrity checks to allow comparison with the previous Sigafoos et al. (2009) review. Studies that used an experimental design (e.g., ABAB, multiple-baseline, control-group design) were rated as having the potential to provide conclusive evidence. Studies that did not use an experimental design (e.g., narrative case studies, pre-post testing without a control group, and A–B designs) were rated as providing inconclusive evidence. Studies with experimental designs also had to have adequate reliability (e.g., inter-observer or inter-rater agreement) and procedural integrity data (i.e., at least 20% of the sessions and 80% or better agreement/fidelity). If a study used an experimental design, but was found lacking with respect to reliability and/or procedural integrity, the certainty of evidence was considered suggestive.

Results

Table 1 provides a summary of each included study’s (a) participants, (b) dependent variables related to communication, (c) intervention characteristics, (d) main intervention outcomes, and (e) certainty of evidence. The sections that follow provide collective summaries of the characteristics for this cohort of 16 studies.

Table 1. Summary of included studies.

Study	Participants	Dependent variable(s)	Intervention characteristics	Main outcomes	Certainty of evidence
1. Ainsworth et al. (2016)	One 14-year-old girl (Anna) with Rett syndrome. Type or stage of Rett syndrome was not specified. The study had eight children (11–16 years) in total, but only had Rett syndrome.	Acquisition of letter-sound correspondence, defined as matching the sound that a letter makes with its orthographic representation.	Intervention was based on the Accessible Literacy Learning (ALL) curriculum (Light & McNaughton, 2009). Anna was taught to look at the printed letters (a, m, p, t, c) and select the letter that matched a sound made by the researcher (e.g., Show me the letter that makes the sound mmmm.). She did this using eye gaze with or without reach approximation. The four-step instructional protocol involved (a) presenting a letter and the sound it made, (b) modeling the correct response, (c) repeating the sound while delaying the model and adding additional foils, and (d) presenting a sound and asking the participant to select the correct letter from a four-choice letter array. Data were collected during tests that occurred prior to each intervention session. Tests involved one presentation of each of the five sounds. Anna received eight teaching sessions on the five letter sounds.	Anna’s mean number of correct responses increased from 1.2 in baseline to 3.0 in the intervention phase.	The certainty of evidence is conclusive because the intervention was evaluated using a multiple-baseline across groups design. Inter-observer agreement and procedural integrity data were adequate.
2. Bartolotta and Remshifski (2013)	Four girls with classic Rett syndrome aged 5, 10, 11, and 14 years. All were at Stage III. None had any spoken language.	Total number of student communication bids directed at an adult and total number of student responses to an adult’s communication bids. Bids and responses could include eye gaze, gestures, vocalizations, body movements, picture symbols, or AAC switch.	During mealtimes, adults were taught to use four strategies to increase student bids and responses. (a) create choice making opportunities, (b) presentAAC devices, (c) wait for child initiation, and (d) respond to child attempts.	Student bids and adult responses increased for all four dyads from Phase 1 (baseline) to Phases 2 and 3 (coaching and post-coaching). Responses decreased at Phase 4 (maintenance).	The certainty of evidence is inconclusive due to the pre-experimental ABCD design and lack of procedural integrity checks. However, adults showed increased bids and responses after baseline, which suggests they were using the targeted strategies.
3. Byiers et al. (2014)	Three females with classic Rett syndrome aged 15, 27, and 47 years. Stage was not specified. MECP2 mutation was confirmed in one participant. All were nonverbal.	Touching a voice-output switch to request preferred stimuli (e.g., TV, attention, or massage). Rate of switch activations per min was calculated for each 5-min session.	A prior assessment indicated that prelinguistic acts (whining, self-injury, or grabbing) were requests for TV, attention, or massage. Intervention aimed to replace these acts with activation of a voice-output switch. Switch use was prompted using most-to-least hierarchy and reinforced with praise and the preferred stimulus. Both independent and prompted responses were reinforced.	When access to preferred stimuli was contingent on prelinguistic acts (i.e., in the baseline or A phases), prelinguistic acts occurred at higher rates than switch activations. When access to preferred stimuli was contingent on switch activations (i.e., in the intervention or B phases), switch activations occurred at higher rates than prelinguistic acts.	The certainty of evidence is conclusive due to the ABAB design and adequate inter-observer agreement and procedural integrity data.
4. Chou et al. (2019)	23 females with Rett syndrome with a mean age of 14 years (range = 3–20). Type and stage of Rett syndrome was not specified, but all had confirmed MECP2 mutation.	The main dependent variables related to communication were participants’ scores on the Receptive, Expressive, and Written Communication sub-domains of the Vineland Adaptive Behavior Scales (Sparrow et al., 1984). Clinical ratings on language and non-verbal communication were also collected.	A group-based music therapy program was implemented with 11 participants. The other 12 served as the control group. Sessions were conducted by a certified therapist with a caregiver (usually the mother) present. Sessions lasted 2 h and occurred twice per week for 24 weeks. Therapy involved vocals and instruments that participants could listen to, look at, and touch. The therapist and caregivers also talked, sang, hummed, and played instruments. Instrument played varied (e.g., stop/start) to encourage participants to follow along.	Participants who received music therapy showed significant increases in their Vineland Receptive sub-domain scores. Clinical ratings of non-verbal communication and language functioning also improved from pretest to post-test for the music therapy group.	The certainty of evidence is inconclusive because participants were not randomly assigned to groups. However, pretest results indicated the two groups had comparable Vineland scores and clinical ratings of communication ability at the start of the study. Inter-rater agreement and procedural integrity do not seem to have been assessed.
5. Fabio et al. (2013)	One, 21-year-old female with Rett syndrome and confirmed MECP2 mutation. She began intervention at 8 years old and was classified as being in Stage IV. She could not speak any meaningful words.	The dependent variable was selecting the correct symbol (e.g., picture, drawing, or printed word, syllable, or letter) from a two-choice array when presented with a corresponding sample stimulus.	Intervention was based on Feuerstein et al. (1988) mediated learning approach, which induces the individual to classify, compare, group, label and convey meaning. Instruction included stating rules reinforcement, interrupting stereotypy, modeling, shaping, fading, and prompting the person to attend. The participant first received pre-training on basic concepts and skills (e.g., staying seated, making choices in a match-to-sample paradigm, and recognizing objects). Pre-training involved 4, 50-min sessions per week from age 8 to 11. She then received discrimination training on matching spoken words to pictures or printed words, matching spoken syllables and letters to printed syllables and letters, selecting printed words to form two-word sentences and responding to questions. This involved 4, 50-min sessions per week from age 11 to 21.	The participant learned 16 words in the matching spoken word-to-printed-word task. She also learned 18 syllables and 19 letters.	The certainty of evidence is inconclusive due to lack of an experimental design. The study involved a B-only (i.e., treatment-only) design. Reliability of data collection and procedural integrity does not seem to have been assessed.
6. Fabio et al. (2018)	Three females with Rett syndrome, aged 29, 30, and 31 years. Type of Rett syndrome was not specified, but all had confirmed MECP2 mutation and all were said to have gone through all four stages. All three were reported to produce sounds. One participant used two words (home and cake) to request.	Dependent variables were (a) producing the correct word or initial letter sound in response to being shown a picture of an object, (b) looking at or touching the correct picture choice, from a two-choice array, when presented with a spoken word instruction (e.g., Touch the duck.). Alpha, beta, and theta bands were recorded using EEG signals.	Intervention involved transcranial direct current stimulation and cognitive empowerment training. Transcranial direct current stimulation occurred 20 min per day over a 10-day training phase. Cognitive empowerment training consisted of prompting and imitation and generalization training.	Intervention was associated with increases in correct words and sounds produced and identified. Alpha beta, and theta bands also increased.	The certainty of evidence is inconclusive due to lack of an experimental design. The design consisted of four sequential phases: Pretest, Training, Post-test, and Follow-up. Inter-rater agreement on utterances was good, but procedural integrity on implementation of cognitive empowerment training does not seem to have been assessed.
7. Fabio et al. (2020)	Thirty-one females with Rett syndrome from 13 to 35 years of age. Mean age was 18 years. All had MECP2 mutation and were in Stages III or IV.	The dependent variables related to communication were production of vowels, phonemes and words in response to picture stimuli.	Participants were randomly allocated to a cognitive empowerment training program with either concurrent sham versus non-sham transcranial direct current stimulation. Participants received 10 daily intervention sessions for 1 week. Sessions lasted 20 min. In the sham condition, 1 mA of stimulation was provided for 30 s, which did not affect neurons. In the non-sham condition, 2 mA of stimulation was provided, Cognitive empowerment training consisted of prompting and imitation and generalization training.	Participants in the non-sham group produced a higher number of vowels and phonemes at post-testing and at the 1-month follow-up. No differences were found for number of words produced.	Certainty of evidence is suggestive due to randomized two-group experimental design. However, checks on reliability of data collection and procedural integrity do not seem to have been undertaken.
8. Fabio et al. (2021)	Twenty-eight females with Rett syndrome from 4 to 22 years of age. Mean age was 12 years. All had MECP2 mutation and were in Stages III or IV.	The dependent variables related to communication were production of correct vowels, phonemes and words in response to picture stimuli.	Cognitive rehabilitation software was configured for use with a Tobii Series-1 eye-tracker. The software presented a target stimulus (e.g., pictures of objects, colors, or shapes) and a distractor. Participants were instructed to give the corresponding word to target stimuli. Correct responses were reinforced with praise. Errors were corrected using verbal prompting and modeling (e.g., What is this? This is bee.). Participants received 30 min of cognitive rehabilitation, 3 days per week for 1 month.	Intervention did not produce a significant effect on the production of correct vowels, phonemes, or words.	Certainty of evidence is suggestive due to the ABAB-ABA withdrawal design. However, checks on reliability of data collection and procedural integrity do not seem to have been undertaken.
9. Hackett et al. (2013)	One, 4-year-old girl with Rett syndrome and confirmed MECP2 mutation. Type and stage of Rett syndrome was not specified. No details were provided on entry-level communication ability	The frequency of turn-taking responses was retrospectively analyzed from videotape. Turn taking included interacting with the therapist via vocalizing, eye gaze, or chest-patting, which were considered to have communicative intent.	14 individual, music therapy sessions were videotaped over a 6-month period. Sessions lasted up to 30 min and were conducted by a registered music therapist. Sessions included a hello and goodbye song and improvization based on the Alvin model (Bruscia, 1987). For example, the participant could direct the therapist’s actions and make choices via eye gaze.	Frequency of turn talking showed a gradual increase over the 14 sessions.	Certainty of evidence is inconclusive due to lack of an experimental design. The study used a B-only (i.e., treatment-only) design. Inter-rater agreement on the scoring of videotapes was good, but procedural integrity data were not collected.
10. Hsieh et al. (2021)	The sample of six children/adolescents (4–19 years) included two girls (Sarah and Anne) with Rett syndrome (aged 4 and 17 years). Type or stage of Rett syndrome was not specified.	Dyadic communication behaviors were coded from 12 video clips obtained during interactions with an adult communication partner. Dyadic communication behaviors included turns and moves (e.g., initiating a conversation and responding to an initiation). Communication functions (e.g., requests for attention and comments) were coded as was stage of communication (e.g., pre-intentional or intentional).	Intervention involved use of eye-gaze assistive technology to control a computer-based communication interface. Naturalistic communication interactions with an adult partner occurred under two conditions: with versus without eye-gaze technology. The eye-gaze technology was the Tobii I-series or PCEye Mini. The communication software was the Communicator 5, Grid 3, or Communicator with WordPower. Communication devices included from 3 to 62 symbols per page. In the non-eye-gaze technology condition, participants used communication boards, eye-gaze frames or iPads.	Significantly more communication initiations were made by the children under the eye-gaze assistive technology condition. Sarah and Anne made more requests for information under the eye-gaze technology condition. They also used the technology more often than gestures in the technology condition.	The certainty of evidence is inconclusive because it is not clear whether or not an experimental design was used. Inter-coder agreement for the video analysis was adequate, but procedural integrity data were not reported.
11. Simacek et al. (2016)	Two females with Rett syndrome, aged 27 and 7 years. The adult had an atypical variant and the child had classic type. Genetic testing had not been completed with the adult. The child had a confirmed MECP2 mutation. Stage of Rett syndrome was not specified. Neither produced any spoken words.	The dependent variable was independent selection of target symbols (photographs of preferred items) from the display of a speech-generating device. The adult’s device had three pages. Each page had a target symbol and three distractors. She selected the target symbol by touching. The child used a Tobii eye tracking device and selected target symbols using an eye gaze response. The Tobii was configured with three pages, each containing a target symbol and one distractor.	Intervention sessions consisted of 5–10 requesting opportunities and lasted 5–15 min. Three to six sessions were conducted per day and 1–2 times per week. Opportunities were initiated by briefly removing a preferred stimulus (e.g., beads or TV) and instructing the participant to Tell me what you want. If the target request did not occur within 1 min, responses were prompted using most-to-least prompting and other stimulus prompting and fading strategies. Correct symbol selections were reinforced by access to the requested item.	Accurate and independent selection of the target symbols increased to 90–100% for the adult and for two of the three items for the child. The child’s request for one item (food) also increased, but performance was more variable, ranging from 14% to 100% over the final 15 intervention sessions.	The certainty of evidence is conclusive due to the multiple baseline across target symbols design and there was adequate inter-observer agreement and procedural integrity data.
12. Simacek et al. (2017)	Three girls (3.5–4 years), but only one (Sidney) had Rett syndrome. Details on mutation, type and stage of Rett syndrome were not provided. Her speech was limited to babbled sounds (ma-ma and ba-ba).	The dependent variable was frequency of requesting responses in three contexts. The response involved activating a BIGmack switch with voice output “more.” The three contexts were requesting food when seated in a high chair, requesting snacks when in the living room, and requesting parent attention when seated on the living room couch.	Parents were taught to use Functional Communication Training (FCT) via telehealth. FCT involved most-to-least prompting of the BIGmack switch and reinforcement of each response by delivering requested item. Parent training involved summarizing the instructional steps and modeling the most-to-least prompt procedure. Up to seven 5-min sessions were conducted per day.	With intervention, use of the BIGmack switch increased and prelinguistic acts (e.g., reach, hit tray) decreased for two of the three targeted requesting contexts (i.e., requests for food and requests for snacks). There was slightly more use of the BIGmack switch than prelinguistic approaching for the requesting attention context.	The certainty of evidence is conclusive due to an adapted multiple probe across requests design and adequate inter-observer agreement and procedural integrity data.
13. Stasolla et al. (2014)	Three girls (8–10 years of age) with Rett syndrome. Details on mutation, type and stage were not provided. All three lacked speech, but made some vocalizations.	Number of items requested by touching photographs or touching pictures on a speech-generating device. The speech-generating device produced relevant voice output (e.g., “orange juice”).	Intervention sessions were 10 min. Sessions occurred 2–4 times per day and 5 days per week over a 4-month period. Participants were taught to make requests for preferred items with a variation of the Picture Exchange Communication System (PECS, Phases 1–3; Bondy & Frost, 2001) and by using a speech-generating device. Participants were prompted to make requests using physical guidance. Requests were reinforced by providing access to the requested item.	The number of requests made per session increased from 0 to 2 in baseline to 5–10 during intervention. The number of requests made per session was comparable under the two conditions (i.e., PECS versus speech-generating device).	The certainty of evidence is suggestive due to the modified alternating treatments with imbedded withdrawal design and adequate inter-observer agreement. However, procedural integrity checks were not reported.
14. Titlestad and Eldevik (2019)	Two (twin) girls (3 years, 7 months old) with Rett syndrome and confirmed MECP mutation. They were at Stage II when intervention began and progressed to Stage III over the course of intervention. Type of Rett syndrome was not specified. Parents reported the girls used vocalizations, eye gaze, and facial expression.	The dependent variables related to communication were participants’ raw scores and age equivalencies for the Receptive, Expressive, and Written Communication sub-domains on the Vineland Adaptive Behavior Scales (Sparrow et al., 2005) and scores from the Assessment of Basic Language and Learning Skills-Revised (ABLLS-R; Partington, 2008).	Starting at 3 years, 7 months of age, the girls received Early Intensive Behavioral Intervention. They received 30–35 h of intervention per week over a period of 33 months. Preschool staff implemented an intervention protocol based on Eldevik et al. (2010) and Smith et al. (1995). These protocols involve a range of strategies (e.g., shaping, prompting, and reinforcement). Teaching goals were derived from the ABLLS-R. Data were collected at intake and after 1, 2 and 3 yrs of intervention.	Vineland Communication scores increased from 31 to 33 points by the end of the intervention. Gains for sub-domains were 14–15 points (Receptive) and 17–18 points (Expressive). Communication). Gains were also evident on the ABLLS-R.	Certainty of evidence is inconclusive due to lack of an experimental design. The study used a B-only (i.e., treatment-only) design, Inter-rater agreement does not seem to have been undertaken, but staff achieved 90% on a treatment fidelity checklist.
15. Vessoyan et al. (2018)	Four girls with Rett syndrome, aged 9, 10, 15, and 15 years. All were at Stage III, but type and mutation confirmation were not reported. Participants used eye gaze, vocalizations, and body movements.	The extent to which participants achieved individualized communication goals (e.g., select messages from AAC device to request item and ask questions) was rated using Goal Attainment Scaling (Kiresuk et al. (1994).	Participants received eye-tracking technology for using speech-generating devices. The devices were configured with from 8 to 30 locations per page. Therapists made five home visits over a 12-month period. During visits, parent and therapist developed communication goals and parents received consultation on use of the eye-tracking technology. They were also consulted on how to promote communication by (a) presuming competence, (b) following the child’s lead, (c) acknowledging responses, and (d) assigning meaning to communication attempts.	All participants were rated as having met or exceeded the expected level of performance for at least one of their two communication goals.	Certainty of evidence is inconclusive due to lack of an experimental design. The study used an A-B-design, but the baseline was derived from retrospective perceptions of participants’ pre-intervention levels of performance. Inter-rater agreement checks were made on goal ratings, but agreement across independent raters was only 50% Procedural integrity does not seem to have been assessed.
16. Wandin et al. (2021)	Three women with Rett syndrome, aged 27, 29, and 31 years. Two had a confirmed MECP2 mutation. Stage and type were not specified. None had any speech.	The dependent variables related to communication were participants’ (a) rate of synthesized words per min, and (b) rate of unique synthesized words per min. Unique words were new words not seen at (a)	Participants were fitted with a gaze-controlled AAC device (Tobii l12+) with communication software (Communicator 5). The AAC devices were configured with communication symbols, such as “MORE”, “FINISHED” and “GOOD.” Participants received 13–16 intervention sessions each lasting 20–30 min over a 6- week period. Intervention involved aided language modeling (Sennott et al., 2016), which involved having the communication partner point to symbols on the AAC device twice per min while speaking.	Moderate to large effects were reported for two participants on both dependent measures. There was no intervention effect for the third participant on synthesized words per min, but a large effect for unique synthesized words.	Certainty of evidence is suggestive due to the ABABA (modified withdrawal) design. Inter-observer agreement was adequate. Formal procedural integrity checks were not conducted, but daily feedback was undertaken with the interventionist to ensure that key aspects of the procedure were implemented correctly.

Note. BIGmack^®, AbleNet, Inc. Roseville, MN.

Participants

Across the 16 included studies, 100 participants with Rett syndrome received intervention. Another 12 participants with Rett syndrome served as a control group (Study 4). The sample size of individual studies ranged from 1 to 31 (M = 7 participants per study). Four studies included only one participant (Studies 1, 5, 9, and 12). Nine studies included two to four participants (Studies 2, 3, 6, 10, 11, 13–16). The three remaining studies (Studies 4, 7, 8) were comparatively large, with sample sizes of 23, 31, and 28 participants, respectively.

All 112 individuals were female ranging from 3 to 47 years of age. Precise ages for individual participants were available for 30 participants because the three studies with the largest sample sizes (Studies 4, 7, 8) provided only age ranges and means. The overall grand mean age can be estimated at 14.5 years. Eight studies included at least one adult, 18 years of age or over (Studies 3–8, 11, 16). Seven studies included at least one preschool-aged child, 5 years or younger (Studies 2, 4, 8–10, 12, 14). Ten studies included at least one school-aged child/adolescent from 6 to 17 years of age (Studies 1–4, 7, 8, 10, 11, 13, 15).

Confirmed MECP2 mutations were reported for 92 participants included in Studies 4–9, 14 and 16 and for one of the two participants in Study 11; thus, 83% of the 112 total participants had a confirmed MECP2 mutation. Type of Rett syndrome (classic or atypical) was reported in two studies (Studies 2 and 11) representing six participants. Five had the classic type and one had an atypical variant. Participants’ stage of Rett syndrome was reported in seven of the 16 studies (44%). These seven studies (Studies 2, 5–8, 14, 15) included 73 participants, representing 65% of the total participant pool of 112. All but two of these 73 participants were in Stage III or IV. The other two participants, from Study 14, transitioned from Stage II to III during the intervention.

The presence or absence of functional speech (e.g., using spoken words to communicate) was reported in 12 studies covering 85 participants (Studies 2, 3, 5–8, 11–16). Only 1 of these 85 participants had functional speech (i.e., the participant used two words to make requests). For the remaining 27 participants across four studies (Studies 1, 4, 9, 10), information on their entry level spoken language abilities was not provided.

Dependent variables related to communication

Five different types of communication behavior served as dependent variables across the studies. Five studies (Studies 3, 11–13, 15) recorded the frequency with which participants selected graphic symbols or activated microswitches/speech-generating devices to request preferred items or events, such as requesting orange juice or asking to watch television. In Study 15, participants could also have other individualized goals that served as dependent variables, such as selecting messages from an AAC device to ask questions. In five other studies (Studies 1, 5–8), performance on various match-to-sample tasks was the communication behavior of interest. For example, selecting the correct choice picture or printed word from a two-choice array when presented with the corresponding spoken-word sample. Three studies (Studies 2, 9–10) recorded general conversational/turn-taking responses, such as initiating a communication interaction and responding to the communication initiations of a non-disabled adult communication partner. For two studies (Studies 4 and 14), standardized assessment scores served as dependent variables. Both of these studies reported scores from the Receptive, Expressive, and Written sub-domains of the Vineland Adaptive Behavior Scales (Sparrow et al., 1984, 2005). Study 14 also used the Assessment of Basic Language and Learning Skills-Revised (ABLLS-R; Partington, 2008) as a dependent measure. Finally, Study 16 recorded the rate at which participants selected core words (e.g., “MORE,” “FINISHED,” and “GOOD”) using a gaze-controlled communication device. Once a word had been selected, the device produced corresponding synthesized speech output.

Intervention characteristics

Seven different types of intervention approaches were implemented across the studies. These were (a) systematic instruction based on principles of applied behavior analysis (ABA), (b) staff/parent training, (c) music therapy, (d) implementation of specific curriculum, (e) aided language modeling, (f) transcranial direct current stimulation, and (g) a naturalistic interaction approach.

Eight studies (Studies 3, 6–8, 11–14) involved the use of various ABA-based systematic instructional tactics, such as most-to-least prompting, modeling, and reinforcement. Some of these studies employed named interventions, specifically: Functional Communication Training (Carr & Durand, 1985; Studies 3, 12), Early Intensive Behavioral Intervention (Reichow et al., 2014; Study 14), and a variation of the Picture Exchange Communication System (Bondy & Frost, 2001; Study 13). Three of these studies employed approaches referred to as Cognitive Empowerment Training (Studies 6, 7) and Cognitive Rehabilitation (Study 8), but the actual teaching procedures (e.g., verbal prompting, modeling, imitation, and generalization training) were based on operant conditioning/ABA principles.

The next most frequently used approach (used in three studies) involved training or coaching adult staff or parents. Study 2, for example, taught adults to use four strategies (i.e., create opportunities, present AAC devices, wait for child initiation, and respond to prelinguistic acts). Adults were taught to use these strategies to increase the child’s communicative initiations and responses. Study 12, already referenced above as an example of an ABA-based intervention, taught parents to use Functional Communication Training. Study 15, coached parents to use four strategies (a) presume competence, (b) follow the child’s lead, (c) acknowledge child responses, and (d) assign meaning to the child’s communicative attempts. These strategies were aimed at enabling participants to activate speech-generating devices via eye-tracking technology.

Two studies used music therapy (Studies 4, 9) and two studies implemented a specific curriculum; specifically the Accessible Literacy Learning (ALL) curriculum (Light & McNaughton, 2009) (Study 1), and Feuerstein et al. (1988) mediated learning approach (Study 5). These curriculum-based programs appear to have also incorporated various systematic instructional procedures (e.g., time delay, modeling, shaping, and fading) to teach specifically sequenced skills. Another two studies (Studies 6 and 7) used transcranial direct current stimulation. Study 16 used an aided language modeling intervention (Sennott et al., 2016). With this approach, the communication partner pointed to symbols on the participant’s AAC device to model the use of the symbols for communication purposes. Finally, the intervention used in Study 10 was described as a type of naturalistic communication interaction. Interactions were conducted with and without eye-gaze technology to determine if that technology would facilitate the participant’s use of a computer-based communication device.

Intervention intensity was defined as the total number of minutes or hours of intervention received by participants, which provides a rough estimate of one common measure of treatment intensity (Warren et al., 2007). Calculations were derived from multiplying the number of individual intervention sessions by session duration and/or the overall length of the intervention phase. The most intense interventions were used in Studies 5 and 14. The 21-year-old participant in Study 5, for example, received 4, 50-min sessions per week over a 10-year period resulting in approximately 1733 h of intervention. The twins participating in Study 14 each received from 30 to 35 h of intervention per week for 3 years for a total of 4680–5460 h of intervention. The next most intense intervention was delivered to the 23 participants in Study 4, each of whom received 96 h of music therapy. In Study 13 participants received from approximately 26–53 h of intervention, which was aimed at teaching requesting skills using picture exchange and speech-generating devices. In half of the studies (Studies 1, 3, 6–9, 11–12, 16) the amount of intervention per participant ranged from about 1 to 10 h. Intervention intensity could not be determined for three studies (Studies 2, 10, 15).

In terms of communication modalities, half of the included studies used one or more forms of aided AAC (Studies 2, 3, 10–13, 15, 16). Aided AAC forms included (a) pointing to photographs or picture symbols (e.g., Studies 2 and 13), (b) touching a microswitch that then activated speech output (e.g., Studies 2 and 3), and (c) selecting symbols on a computer-based speech-generating device either directly or via eye-gaze technology (Studies 10, 11, 13, 15, 16). Seven studies targeted receptive and/or expressive speech modes (Studies 1, 4–8, 14), whereas Study 9 focused primarily on increasing prelinguistic behavior (e.g., body movements, eye gaze, and vocalizations) for some of the participants.

Intervention outcomes

Thirteen studies (Studies 1–3, 5, 6, 9–16) reported positive intervention outcomes. Positive outcomes included, for example: (a) increases in the percentage of correct responses (Studies 1, 11), (b) increases in the number of words, syllables, or letters learned (Studies 5, 6), (c) increased use of AAC devices to make requests (Studies 3, 12, 13), (d) increased frequency of communication initiations and responses to initiations (Studies 2, 9, 10), (e) increased Vineland scores in the communication domain (Study 14), and (f) attainment of individually set communication goals (Study 15). Two studies reported mixed outcomes (Studies 4 and 7). Specifically, in Study 4, music therapy was associated with significant increases in Vineland scores, but only for the receptive communication sub-domain. In Study 7, there was an increase in the production of vowels and phonemes for participants in the non-sham transcranial direct current stimulation condition, but no significant differences in the sham versus non-sham conditions on the number of words produced. Only one study (Study 8) reported a lack of any intervention effect for communication-related variables. In that study, intervention did not increase the production of correct identification of vowels, phonemes, or words.

Certainty of evidence

The certainty of evidence was rated as conclusive for four studies (Studies 1, 3, 11, 12). All four of these studies employed experimental designs and included adequate reliability (e.g., inter-observer or inter-rater agreement) and procedural integrity data. The evidence from another four studies was rated as suggestive (Studies 7, 8, 13, 16). These four studies employed experimental designs but lacked formal reliability and/or procedural integrity checks. The remaining eight studies (Studies 2, 4–6, 9, 10, 14, 15) were rated as providing inconclusive evidence because of pre-experimental designs and lack of formal reliability and/or procedural integrity checks.

Discussion

A previous review (Sigafoos et al., 2009) identified nine studies, published over a 10-year period (1995–2005), on enhancing communication functioning of people with Rett syndrome. The present updated review—covering the 12 year-period from 2010 to May 2022—identified 16 new studies. These 16 new studies extend the literature in relation to participant and intervention characteristics and the overall certainty of available evidence. Given the intervention challenges associated with Rett syndrome, these extensions to the literature could be seen as representing important advances.

The participant pool in this updated review is larger, more age-diverse, and generally better described than for the 2009 review. In terms of the total number of participants, the 2009 review covered 31 participants, whereas the present updated review covers 112 participants, 100 of whom received intervention; thus, the average sample size of studies increased from just over three participants per study in the 2009 review to a mean of seven participants per study in this updated review.

In addition to increased numbers of participants receiving intervention, the present set of 16 studies also covered a wider age range. Specifically, participant ages in the 2009 review ranged from 2:7 to 17:0 (years:months), whereas the participants covered by the current updated review ranged from 3 to 47 years of age. Eight studies in the present review included at least one adult participant, aged 18 years or older. The inclusion of adult participants is a significant extension given the relative lack of research into their learning and behavioral characteristics (Matson et al., 2008). Encouragingly, all but one of the eight studies that included adult participants (i.e., all but Study 8) reported positive outcomes. These generally positive outcomes support the potential value of providing communication intervention to the adult population, which is significant in light of recent evidence showing increased medical problems and deterioration of mood in people with Rett syndrome as they age (Cianfaglione et al., 2016). Such age-related problems might be expected to complicate the challenges already inherent in providing effective communication intervention to people with Rett syndrome. While it is encouraging that seven studies showed some positive intervention outcomes for adults with Rett syndrome, it is important to note that the oldest participant was 47 years old; thus, the extent to which even older adults might show a positive response to intervention will require additional research. It is also the case that none of the studies included in this review examined the potential impact of age-related health and mood problems. Exploring response to intervention in light of age-related changes would be a useful future research direction.

In the present review, most participants had a confirmed MECP2 mutation (83%). No such confirmation was provided in any of the nine studies included in the 2009 review, which is understandable given that the genetic basis of Rett syndrome was not discovered until 1999 (Amir et al., 1999). In addition, genetic testing in neurodevelopmental disorders has only recently become more routine (Du et al., 2022). The currently high confirmation rate increases diagnostic validity (Smith et al., 2003), which is critical for comparing response to intervention across and within different populations, such as comparing children with Rett syndrome to children with Angelman syndrome.

Similar to the results of the 2009 review, participants were mainly in Stage III or IV, and had the classic variant; however, the stage and type of Rett syndrome were not consistently reported. Also, none of the studies in either the present update or the 2009 review was designed to compare intervention outcomes for participants in different stages or variants. These could be important variables to isolate in future research. It is possible, for example, that individuals with the preserved speech variant of Rett syndrome (Skjeldal, 2020) might be more responsive to speech-mode interventions.

Only one study in the current updated review provided intervention to participants who were reported to be at Stage II of the disorder (Study 14). The twin girls in that study entered intervention at Stage II and progressed to Stage III over the course of a 3-year early intensive behavioral intervention program. Stage II is associated with rapid developmental regression of acquired motor, communication, and social skills (Fabio et al., 2020). In this second stage of Rett syndrome, children also tend to lose purposeful hand function. Stage II changes would likely have an impact on the child’s response to intervention, including any intervention aimed at enhancing communicative functioning. In contrast, while Stages III and IV are associated with severely restricted adaptive behavior functioning, these stages also seem to be associated with less rapid skill loss, which would seem to be more conducive to intervention. This could also be why most communication intervention research has involved participants in these later stages. However, it is also important to note that the developmental progression and characteristics associated with Rett syndrome may not necessarily follow or conform to a standard series of four distinct stages (Feldman et al., 2016). There appears to be little research on the impact of intervention on the developmental trajectory of individuals with Rett syndrome. An interesting line for future research would be to investigate whether it is possible to preserve more speech in persons with atypical Rett syndrome through early intensive speech therapy.

Implications for practice

The present updated review provides additional empirical support for the use of systematic instruction procedures. Several studies in both the present updated review (e.g., Byiers et al., 2014; Simacek et al., 2016; Stasolla et al., 2014) and the previous 2009 review (e.g., Sigafoos et al., 1996; Van Acker & Grant 1995) demonstrated successful acquisition of aided AAC for functional requesting via the application of various ABA-based systematic instructional tactics. This is not surprising given that such procedures are well established for teaching a range of adaptive behaviors to individuals with neurodevelopmental disorders (Storey, 2022). The effectiveness of these procedures for teaching communication skills to individuals with Rett syndrome may depend on the extent to which the targeted response mode (e.g., touching a switch or selecting graphic symbols) is compatible with participants’ motor abilities. Acquisition might also depend on the extent to which participants are motivated to communicate (Drasgow et al., 1999). There is some consensus in the AAC field that teaching participants to request highly preferred stimuli is one way to help ensure or increase motivation and might thus represent a useful initial communication function to target when beginning an AAC intervention (Still et al., 2014).

Three studies employed intervention approaches that appear to be widely applied in practice, specifically music therapy and aided language modeling. Music therapy was used in two studies (Studies 4 and 9) with modestly positive results reported for the dependent variables related to communication; however, the specific music therapy procedures that were applied differed across these two studies as described in Table 1. To date, there are not enough data on any one approach to the provision of music therapy for enhancing communication in Rett syndrome. In terms of future evaluations, it is generally the case that music therapy tends to have a more comprehensive focus. Consequently, it could have varying effects on different areas of functioning (e.g., social interaction versus communicating a request). In light of this, a systematic review of the wider music therapy literature involving individuals with Rett syndrome would seem timely. The one study (Study 16) using aided language modeling reported positive outcomes in terms of increased production of synthesized speech output. While several studies have reported positive effects of this approach with other populations (Sennott et al., 2016), additional studies are needed to ascertain the extent to which such approaches might also be effective for improving the communication functioning among people with Rett syndrome more generally.

In addition to these more well-established intervention approaches, a few studies evaluated novel intervention approaches. One such novel approach involved the use of eye tracking technology. This technology appears promising and could provide a means by which participants with significant motor impairment could reliably access high-tech AAC devices. Another novel intervention involved the use of transcranial direct current stimulation. The two studies in which this was applied (Studies 6 and 7), reported some positive impact on the participants’ communication, but these participants also received systematic instruction. Still, additional research on the use of such stimulation would seem warranted.

Intervention intensity did not seem to be a critical variable. While studies with the most intensive interventions (Studies 5 and 14) reported positive results, other studies demonstrated clear evidence that the targeted communication skills had increased after relatively brief intervention (e.g., 1–10 h). Byiers et al. (2014; Study 3), for example, taught three participants with Rett syndrome to use microswitches to request preferred stimuli, instead of resorting to problematic forms of prelinguistic requesting (i.e., self-injury, whining, or grabbing). All three participants learned this new requesting response within about 60 min of systematic instruction. Instead of total hours of training, targeting discrete communication forms and functions and applying systematic instruction seemed to be the more critical variables; however, this conclusion is tentative because none of these studies set out to investigate the effects of varying treatment intensity. In addition, parameters other than total hours of intervention provided should be considered when evaluating intervention intensity, such as comparing intervention outcomes when intervention sessions vary in terms of the number of teaching opportunities provided and responses made (Warren et al., 2007). Reichle et al. (2021) noted “Treatment intensity is an aspect of communication intervention that has received limited empirical attention” (p. 146). Future research on a range of parameters related to treatment intensity could make for more efficient delivery of communication intervention.

Limitations and future directions

Only one of the nine studies in the previous 2009 review (Sigafoos et al., 2009) was judged to provide conclusive evidence. In contrast, eight studies in this updated review (50%) were rated as providing conclusive or suggestive evidence. This change is due to a higher percentage of studies using experimental designs to evaluate intervention effects. In that respect, the methodological quality of the communication intervention literature for people with Rett syndrome could be said to have improved over the past 12 years. This conclusion should be interpreted with caution, however, because in the present review was limited in its approach to the ascertainment of certainty of evidence. Specifically, we only considered whether the study had an experimental design and adequate reliability and procedural integrity data. Other factors, such as replicable procedures and operationally defined dependent variables, are also relevant when judging the certainty of evidence; however, these latter aspects would seem to involve more subjectivity of judgment. Still, this review has identified new studies that collectively add much needed quantity and quality to the intervention literature in enhancing communicative functioning in persons with Rett syndrome.

In terms of future directions, subsequent reviews could aim to include grey literature and other steps to mitigate against publication bias. As mentioned before, future reviews would also be improved by considering a wider range of variables that influence a study’s certainly of evidence. Consideration of a wider range of parameters related to treatment intensity (e.g., number instructional opportunities) would also likely strengthen the quality of future reviews.

Conclusion

This updated review identified 16 studies that collectively would seem to extend the generality or external validity of the literature on communication intervention for people with Rett syndrome. Establishing the generality of intervention research depends in part on participant numbers and their characteristics (Murad et al., 2018). The present set of 16 studies included a relatively large and more age-diverse participant pool and thus extends the overall generality of the literature on enhancing communication functioning of children, adolescents, and adults with Rett syndrome. This updated review also provides additional experimental demonstrations that systematic instructional procedures can be successfully applied to teach functional use of aided AAC to individuals with Rett syndrome. This review also identified a number of novel intervention approaches that would seem worthy of additional research. Eye tracking technology for enabling participants to activate speech-generating devices is one such innovation that would seem to have considerable potential given the motoric impairments associated with Rett syndrome.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Rett Elternhilfe e.v. (Germany) and Rett Syndrome Europe.