Historical Perspective on Tactile Approaches to Speech Therapy

Recently Caroline Bowen on Twitter (@speech_woman) alerted us to a new fad, therapeutic massage. Of course, the therapy techniques being promoted on the website that Caroline linked to and in the workshops offered by the person in question are not new at all. I recognized them immediately as being very very old and the historical basis for other techniques that are in current use such as tactile cueing and PROMPT. The description of therapeutic massage that was offered (I’m sorry, I refuse to link to the site) put me in mind of Charles Van Riper’s brilliant response to the Motokinesthetic Method which I reproduce here for your edification and enjoyment:

“We have previously mentioned the Motokinesthetic Method invented by Edna Hill Young as one of the approaches used in teaching a child with delayed speech to talk. It has also been used in the elimination of misarticulations. Essentially, this method is based upon intensive stimulation; however, the stimulation is not confined to sound alone but to tactile and kinesthetic sensations as well. The therapist, by manipulation and stroking and pressing the child’s face and body as she utters the stimulus syllable, helps him recognize the place of articulation, the direction of movements, the amount of air pressure, and so on. Watching an expert motokinesthetic therapist at work on a lisper is like attending a show put on by a magician. The case lies on a table with the therapist bending over him. First she presses on his abdomen to initiate breathing as she strongly makes the s sound; then to produce a syllable from the patient, her fingers fly swiftly to close his jaws, spread the lips, and tap a front tooth, thereby signaling a narrow groove of the tongue or the focus of the airstream. Then her magical fingers squeeze together to draw out the sibilant hiss as a continuant.

One therapist, when working with a child,  used to “draw out” the s, wind it around the child’s head three times then insert it into her ear, thus insuring that it would be prolonged enough to be felt. Each sound has its own unique set of deft manipulations, and considerable skill is required to administer motokinesthetic therapy effectively. Viewed by the cold eye of the modern speech scientist, many of the motokinesthetic cues seem inappropriate; and a therapist would need sixty fingers and thirty arms to provide sufficient cues to take care of the necessary integration and coarticulation. Moreover, much of our research has indicated that standard sounds are produced in different ways by different people, and that their positioning vary widely with differing phonetic contexts. We suspect that much of the effectiveness of this method is due to its powerful suggestion (the laying on of hands), to its accompanying auditory stimulation, or to the novelty to the situation, which may free the case to try new articulatory patterns. We have used it successfully with some very refractory cases, but we always have felt a bit uncomfortable when doing so, as though we were the Magical Monarch of Mo in the Land of Hocus Pocus. (p. 198-201).”

This is just one of many delightful passages from Van Riper’s book “Speech Correction: Principles and Methods” (1978 Prentice-Hall edition but first published in 1939). Characteristically, this passage shows Van Riper to be far ahead of his time. Tactile approaches to speech therapy just seem to make sense because, as I heard numerous times at the ASHA conference last week, “speech comes from movement”. However, a point I make repeatedly in our book Development Phonological Disorders , I believe that this perspective is subtly backwards. Speech movements are learned through practice. The practice is motivated by the desire to achieve functional goals. Learning involves linking knowledge of the goal with the movements used to achieve the goals. The movements are learned through the process of achieving goals which are phonetic, phonological and ultimately linguistic in nature. In another post I will talk more about the issues with trying to shape those movements articulator-by-articulator.

Still no evidence for PROMPT

Some readers may wonder why Francoise and I did not mention PROMPT therapy even once in our book even though Chapter 10 does discuss interventions for Childhood Apraxia of Speech (CAS). The reason is that we found out, when we were using the very excellent Williams, McLeod & McCauley (2010) book for teaching, that students believe that if a treatment is mentioned in a text book that it must be a “good” treatment. Even when provided with information about levels of evidence and so on, the students could not or would not distinguish between treatments that were and were not evidence based. So when we wrote our book we decided to refuse to mention interventions that were in our view NOT evidence based and PROMPT definitely fell into this category. So, when I noticed an “in press” manuscript at the AJSLP site promising evidence in support of this treatment, I thought, ok, maybe we’ll have to add it to the next edition (getting ahead of myself here I know but I’m an optimist). Anyway, back to PROMPT…

The paper, Dale and Hayden (in press) reports 4 case studies in which probe scores for untrained words related to three goals per child are provided at 9 time intervals: baseline, phase 1, phase 2 and follow-up. There are three base-line probes, 2 in each of two treatment phases, 1 just post-treatment and a second follow-up probe 3 months post-treatment. Two of the children received the “PROMPT” intervention without the tactile prompts during the first phase and the full PROMPT intervention with the prompts in the second phase. The remaining two children received these treatment phases in reverse order. The abstract states “untreated word probes provided evidence for more gain when tactile cues were included”. Here is a table summarizing a rough description the untreated word probes as shown in Figures 1 through 4 (I have shaded periods associated with the full PROMPT intervention):

Dale Hayden figure summary

Some may quibble with my interpretation based on the statistics shown in the tables but I don’t agree that these statistical interpretations are valid given that the design is not a proper single-subject design. It is not a multiple baseline design and the nature of the study ensures that a withdrawal design is not appropriate to the question. The baselines are not stable and the probes are not administered frequently enough for a statistical interpretation. I find nearly all of these single-subject treatment studies in SLP to be extremely difficult to interpret in fact and prefer a single-subject randomization design in which a statistical analysis is fully appropriate as I have described in a previous blog post.

Notwithstanding the questionable interpretation of the findings in this study, I am even more bothered by a strange interpretation of the DIVA model in the introduction where it is stated that “In Guenther’s model, early speech learning relies on sensory feedback from the proprioceptive system, which eventually gives way to the auditory modality as the primary source of feedback information.” This misrepresentation of the model is used as a justification for the tactile cueing system used in PROMPT. I would say that, to the extent that their statement makes any sense at all, it is more or less backward, given this direct quote from Callan et al. (2000): “The speech production system must establish a mapping that is able to move the articulators in a manner to reach learned targets, taking into account the current context of the system. Auditory feedback of self-produced speech may serve as an adaptive signal that could establish a mapping that guides the movements of the articulators in order to reach auditory targets”. I have discussed the role of the auditory target and auditory feedback in the development of speech motor control in an open-access paper in relation to speech sound disorders as well (Shiller, Rvachew, Brosseau-Lapre, 2010). During learning auditory and proprioceptive feedback are crucial; over time feedforward control of speech becomes possible, as I describe in Rvachew and Brosseau-Lapre (2012): “The forward model tracks the vocal tract’s position in auditory planning space without relying on auditory feedback; rather the position of the vocal tract is determined by a prediction based on the known consequences of executing the motor plan combined with actual somatosensory feedback during production of the syllable. This information allows for continuous updating of the feed-forward command as the vocal tract moves to target…”. My concern, based on our current studies of children with CAS, is that interventions that provide too many external cues may interfere with the child’s access to and integration of different sources of feedback during the production of speech movements. I’ll have more to say about this potential problem as we begin to publish this work. In the meantime, I hope that readers will be wary of “single subject designs” that are not actual experiments. I also look forward to talking with people about these issues at ASHA 2013 where I will be presenting both a poster and a seminar on CAS interventions.