Case Study: Differential Effects of Melody and Rhythm in Melodic Intonation Therapy

Article information

Abstract

INTRODUCTION

Evidence has been well accumulated for more than two centuries that a patient with non-fluent, or Broca’s, aphasia secondary to a left hemisphere stroke is able to sing songs and words (e.g., Marina, Pasqualetti, & Carlomagno, 2007; Mills, 1904, Gerstmann, 1964, Stahl, Kotz, Henseler, Turner, & Geyer, 2011). The rationale behind this evidence relies on neuroplasticity that a patient’s intact right hemisphere, known to play an important role in processing steady-state spectral information and slowly changing aspects of the auditory-signal (e.g., prosodic contours and sustained vowels), may stimulate the damaged left hemisphere responsible for integrating auditory information across shorter intervals of time (Naeser & Helm-Estabrooks, 1985; Sparks & Deck, 1994; Zatorre & Belin, 2001; Zipse, Worek, Guarino, & Shattuck-Hufnagel, 2014).

This clinical observation has led to a variety of singing-based therapeutic techniques such as Melodic Intonation Therapy (MIT; Albert, Sparks, & Helm, 1973; Helm-Estabrooks, Nicholas, & Morgan, 1989; Sparks, Helm, & Albert, 1974; Wan, Ruber, Hohmann, & Schlaug, 2010). Techniques of MIT include sequentially training a patient to sing simple and familiar words and phrases with melody attached. While singing, a therapist rhythmically taps the patient’s left hand. In other words, MIT consists of three key elements: (1) melody, (2) rhythmic tapping, and (3) formulaic, or simple and familiar (e.g., “How are you?”) words or phrases.

Although the efficacy of MIT on the improvement of verbalization for patients with Broca’s aphasia has been substantially documented (e.g., Conklyn, Novak, Boissy, Bethoux, & Chemali, 2012; Hurkmans et al., 2012; Schlaug, Marchina, & Norton, 2009), there is an ongoing debate as to which element of the three is the driving force of the efficacy of MIT or singing-based therapeutic approaches (e.g., Boucher, Garcia, Fleurant, & Paradis, 2001; Zipse et al., 2014). The original authors of MIT considered the melody as the key component for stimulating the use of intact right hemisphere (Albert et al., 1973) or supporting the right hemisphere to activate residual left-hemisphere’s function (Sparks et al., 1974). Two case observations of MIT or modified MIT attempted to highlight the crucial role of melody controlling for rhythmic left-hand tapping. Hough (2010) reported that one participant with chronic Broca’s aphasia, who exhibited little progress with MIT due to difficulty with the tapping element of MIT, significantly increased production of short phrases using MIT without rhythmic tapping. Wilson, Parsons, and Reutens (2006) concluded that melody might account for the effect of MIT in a neurologically stable amateur male musician with severe expressive aphasia using a controlled pre- and post-treatment design. Consequently, these results appear to be somewhat confounded with the study participants’ unique characteristics (i.e., difficulty in the tapping element and musical training prior to stroke).

The most common finding in both cross-sectional speech facilitation studies and longitudinal treatment studies that attempt to test melody-only, rhythm-only, and combined conditions is that rhythmic condition may be an essential component of the efficacy of MIT (Boucher et al., 2001; Stahl, Henseler, Turner, Geyer, & Kotz, 2013; Stahl et al., 2011). Boucher et al. (2001) suggested that the experimental condition emphasizing rhythm resulted in significant improvements in two participants’ repetition of target utterances, compared to no or little gain following the treatment condition emphasizing melody. Stahl et al. (2011) concluded that the facilitating effect of MIT has its root in rhythm rather than melody. A relatively large number of participants (n=17) and rigorous study design (e.g., controlling for lip-reading, memory, and motor automaticity) allow the finding to be more credible than other case studies. Findings from a longitudinal experiment including 15 participants also confirmed that melody may not benefit speech recovery over and above rhythmic speech (Stahl et al., 2013).

In terms of formulaic speech, it is presumed that the production of common and familiar utterances (e.g., “Happy birthday to you”) is automatized at the motor level. Preserved automaticity in formulaic phrases for MIT revealed that memory and motor automaticity mediate speech production in patients with Broca’s aphasia (Stahl et al., 2011). In other words, individuals with non-fluent aphasia can repeat or produce formulaic utterances easier than propositional (i.e., non-formulaic, creative) speech, utilizing long-term memory and automaticity. Van Lancker Sidtis (2004) and Stahl et al. (2013) proposed that propositional and formulaic speech rely on different neural pathways. The researchers suggested that a patient with non-fluent aphasia should improve propositional speech, reactivating left hemisphere, and formulaic speech, stimulating intact right hemisphere, respectively.

In summary, researchers attempted to better understand the contribution of individual MIT components to its therapeutic effect. It is suggested that production of formulaic utterances for patients with non-fluent aphasia should be trained rhythmically. Until now the effect of melody on speech production in patients with Broca’s aphasia has been somewhat ambiguous, possibly due to lack of the rigor of prior studies. The current case study was designed to demonstrate differential effects of melody-only and rhythm and left-hand tapping-only conditions on accuracy of repetitions in a patient with severe Broca’s aphasia. Formulaic utterances were used in both conditions given the participant’s severe deficit in propositional speech. Considering the nature of case observational study, it is not possible to eliminate extraneous variables or biases. However, the authors hope to demonstrate the functional relationship between individual MIT elements and its therapeutic effect, carefully controlling for the melody-only and rhythm and left-hand tapping-only conditions.

METHODS

Participant

The participant of the present study was a 77-year-old, monolingual, non-Hispanic, Caucasian male with severe Broca’s aphasia and apraxia of speech enrolled at a university clinic. The participant was a retired civil engineer who suffered from a left hemisphere cerebrovascular accident (CVA) six years prior to the intervention study, characterized with severe expressive aphasia, mild to moderate receptive aphasia, and several oral and verbal apraxia. Immediately after the onset of CVA, the participant had been treated with acute and rehabilitation therapeutic services for one year. After then, he had continued to receive speech and language services at the university clinic until the present intervention study was initiated. Previous intervention services in expressive language and speech focused on the following: (1) improving oral communication skills by improving repetition and object naming, (2) extending length of utterances, (3) improving word retrieval skills, and (4) increasing functional communication skills via use of a speech-generating AAC device. The participant’s wife and previous clinicians reported that a plateau in his verbal communication ability had existed for more than one year prior to the intervention study.

The first author, a second-year graduate clinician at the time of data collection, tested the participant one week prior to the study under direct supervision of the third author, who is a certified and licensed SLP with extensive experience in the diagnosis and treatment of neurogenic disorders in adults, using Boston Diagnostic Aphasia Examination-Third Edition (BDAE-3; Goodglass, Kaplan, & Barresi, 2001). The fourth author, a senior undergraduate student at the time of data collection, videotaped the two testing sessions. A summary of selected test results is presented in Table 1.

Table 1

Results of boston diagnostic aphasia examination-third edition

Variables

The independent variable was the individual component of MIT, with two conditions: (1) melody only and (2) rhythm and left-hand tapping only. A step-by-step description of melody-only condition and rhythm and left-hand tapping-only condition is presented in Table 2. The dependent variable is the percentage accuracy of the participant’s ability to repeat a list of targets (i.e., pre-selected words/phrases). The length and complexity of targets depend on the severity of participant’s deficit and personal functional relevance. Based on the results of BDAE-3 and intensive interview of the participant’s wife and previous clinicians, we selected 250 words/phrases and their Boardmaker^© pictures as the targets. Examples of targets are presented in Table 3. None of the targets were introduced to the participant prior to the study, and low baseline performance was obtained.

Table 2

Descriptions of study phases

Table 3

Examples of Target Words and Phrases

Procedure

Treatment and baseline sessions were administered by the first author under direct supervision of the third author. The fourth author videotaped the entire sessions for subsequent data analysis and reliability/fidelity scoring. A summary of study duration and session sequence is presented in Table 4.

Table 4

Study design

Baseline phase

The baseline phase consists of three sessions. The data for the baseline phase were based on the productions on a repetition task in which the first author produced each target at a normal rate when the Boardmaker^© picture was presented without any other cues. No feedback was provided during the baseline task.

Treatment phases

The present study included two treatment phases, and each treatment phase consisted of five sessions. During the treatment phases, the two conditions of melody only and rhythm and left-hand tapping only were compared. Both conditions emphasized slow production of the targets, not exceeding one syllable per second. The treatment task is a repetition task in which the first author produced each target when the Boardmaker^© picture was presented either with melody only or with rhythm and left-hand tapping only. Both conditions were administered every treatment session, with the order of conditions counterbalanced across sessions (e.g., melody presented first in treatment session one; rhythm and left-hand tapping presented first in treatment session two). A randomly-selected 25 targets were assigned in each treatment session. Ten targets were presented under the melody-only condition, and another ten targets were presented under the rhythm and left-hand tapping-only condition. Five control targets without treatment conditions were entered between two conditions in each treatment session.

Maintenance phase

Each treatment phase was followed by a maintenance phase, which consisted of two sessions. The data for the two maintenance phases were based on productions of a repetition task in which the first author produced each target at a normal rate when the Boardmaker^© picture was presented without any other cues. No feedback was provided during the maintenance task.

Follow-up sessions

To monitor a short-term effect of differential learning, two follow-up sessions were implemented; one session was scheduled one month following maintenance phase two, and the other session was scheduled three months following maintenance phase two. The data for the two follow-up sessions were based on productions on a repetition task in which the first author produced each target at a normal rate when the Boardmaker^© picture was presented without any other cues. No feedback was provided during the follow-up task.

RESULTS

Data are plotted in Figure 1 and summarized in Table 5. In the figure, numbers represent the percent of the participant’s accurate repetitions. Different phases are demarcated by vertical dotted lines. First, visual inspection of the data was conducted as suggested by Kromrey and Foster-Johnson (1996). The participant, in general, demonstrated fluctuating and variable performances of accurate repetitions across sessions. During baseline phrase, the participant consistently demonstrated difficulty repeating targets accurately, ranging from 20 to 33% accuracy of trials. During treatment phase one, the participant distinctively demonstrated a high level of accurate repetitions in session four under melody-only condition. The participant’s accurate repetitions in sessions four and five under melody-only condition were significantly superior to the rest of performances during treatment phase one. Treatment phase two exhibited more stable and consistent repetition patterns than treatment phase one. The participant’s performances under melody-only condition were significantly better than his performances under rhythm and left-hand tapping-only condition in four sessions among five. In session five, the participant accurately repeated targets in equal numbers under both conditions.

Figure 1

Percentage accuracy of the participant’s correct repetitions. Mel, melody-only condition; Rhy, rhythm and tapping-only condition; No Tx, no treatment was provided.

Table 5

Total number of trials and correct repetitions

Next, to measure the differential effect of melody-only and rhythm and left-hand tapping-only conditions, the percentage of nonoverlapping data (PND) was calculated by dividing the number of treatment sessions exceeding the highest data point during baseline sessions and dividing by the total number of treatment sessions (Parker, Hagan-Burke, & Vannest, 2007; Manolov & Solanas, 2009; Scruggs & Mastropieri, 2013; Scruggs, Mastropieri, & Casto, 1987). PND statistics higher than 90% are interpreted as highly effective intervention, 70% to 90% as fairly effective treatment, 50% to 70% as questionable effectiveness, and less than 50% as unreliable treatment. The participant’s highest data point during baseline sessions is represented as a horizontal dotted line in Figure 1. PND statistic for melody-only condition is 7/10=70%, which is fairly effective. PDN statistic for rhythm and left-hand tapping-only condition is 3/10=30%, that is, unreliable treatment. Both follow-up sessions’ data were slightly higher than the participant’s highest data point during baseline sessions.

DISCUSSION

The purpose of the current case study was to explore differential effects of melody-only and rhythm and left-hand tapping-only conditions on the accuracy of repetitions in a patient with severe Broca’s aphasia. Formulaic utterances were used in both conditions given the participant’s severe deficit in propositional speech. The study’s preliminary findings have possibly added to the literature on MIT by examining a functional relationship between individual MIT elements and their effects on the accuracy of repetitions in a patient with severe Broca’s aphasia. By carefully controlling for the melody-only and rhythm and left-hand tapping-only conditions with an alternating treatment design and controls of bias (e.g., counterbalancing strategy of the two conditions, random selection of targets, and two treatment phases), these findings have possibly aimed to advance our understanding of the importance of clinical observational research for the contribution of individual MIT elements.

A major finding from the current observational study was that both visual inspection of data and PND calculation indicated an immediate effect of melody-only condition on the participant’s accuracy of repetitions. The positive effect of rhythm and left-hand tapping-only condition was not observed. During the treatment sessions, the participant exhibited a considerable amount of impulsivity. The participant’s speech was halting overall, but it was frequently observed that he rushed to complete repetition as quickly as possible. This impulsivity was accelerated during the rhythm and left-hand tapping-only condition. This condition caused the participant to move ahead of the examiner’s speed, whereas this behavior did not occur during the melody-only condition. The authors’ observation regarding the participant’s difficulty following the tapping was consistent with rhythm-processing deficits that at least some patients with Broca’s aphasia exhibit. For example, Hough (2010) reported a positive effect of MIT when the tapping element was deleted. Tapping was not presented in this study because the participant found the tapping distracting. The researcher speculated that the patient may have had impaired speech as well as impaired tapping, therefore, tapping interrupted attempts at speech production rather than facilitated them. Zipse et al. (2014) supported Hough’s hypothesis that some patients with aphasia have a particular difficulty processing rhythm as opposed to pitch information. It is suggested that therapists should consider modifying MIT with the tapping omitted when (1) patients with aphasia demonstrate little success with MIT, and (2) rhythm-processing deficit is suspected.

Motivation and mood effect of singing in unison can also be considered when examining the positive effect of melody-only condition on the accuracy of repetitions in the participant. Singing is a natural, pleasing, and stress-free way for patients with non-fluent aphasia to verbalize, and thus it possibly motivates the patients to continue to focus on verbal repetition activities (Merrett, Peretz, & Wilson, 2014; Racette, Bard, & Peretz, 2006). A positive relationship between music therapy and patients’ motivation and mood has been substantially observed in the fields of physical rehabilitation and mental health, indicating that singing may be intrinsically motivating (Merrett et al., 2014). Considering that low mood and lack of motivation are commonly observed in individuals with aphasia (Robinson, 2003), the positive influence of singing on the participant’s motivation and mood might have resulted in the improvement of accuracy of repetitions.

Limitations of the current study should be addressed. First, the results of the study were preliminary because of the selection of the participant. The participant’s fatigue, distraction, and overall weakness resulted in fluctuating and variable performances across phases. These extraneous variables negatively influenced the study’s validity given that only one participant was examined. Another limitation was that each treatment phase only included five sessions, and these were not intensive enough to examine the lasting effect of melody-only condition during the follow-up phase. Due to the university clinic’s summer schedule, the entire study including the testing, baseline, treatment, and maintenance phases had to be completed in six weeks. Although the follow-up data did not indicate the lasting effect of melody-only condition, the participant’s wife and authors anecdotally reported noticeable differences in the participant’s initiation of rehearsed speech at home and the clinic. For example, the participant’s greeting was changed from a head nod and the phrase “okay okay” during the baseline phase, to the phrase “I am good,” in later treatment sessions.

Exploring the differential effect of melody-only and rhythm and left-hand tapping-only conditions should not be viewed as a reductionist approach because clinical observations of individual conditions of MIT enrich our understanding of how MIT works for patients with Broca’s aphasia. Given that some evidence for the positive effect of melody-only condition versus rhythm and left-hand tapping-only condition has been reported, attempts to find a simple and single framework explaining the effects of MIT do not appear to be worthwhile. It is possible that pre-morbid factors (e.g., musicianship), lesion factors (e.g., time since onset and location), and comorbidity factors (e.g., aphasia only and aphasia with apraxia of speech) are intertwined to determine the progress of verbalization when MIT is implemented (Merrett et al., 2014). The current case study possibly attempted to extend the literature on the functional relationship between the participant’s pre-morbid, lesion, and comorbidity factors and the positive effect of individual MIT elements on the accuracy of verbal repetitions.

Article information Continued