The SAILS approach to therapy is the result of more than a decade of continuous
research. An early study established that many misarticulating children display
a concurrent receptive auditory difficulty [1]. The first studies used a continuum
of seven synthetic speech stimuli that varied from "seat" to "sheet".
Stimulus 1 was modelled on adult tokens of "seat". The remaining stimuli were
created to be progressively less "seat" like, so that stimulus 4 was highly ambiguous
and stimulus 7 sounded clearly like "sheet". Adults, normally developing children,
and children with delayed phonological skills were asked to listen to the stimuli,
presented in random order, and indicate whether they heard "seat" or "sheet".
Normally developing children and about a third of the speech delayed children
produced identification functions like those of adults. They were easily able
to categorize the stimuli into two groups, with the first 3 or 4 stimuli heard
as "seat" and the last 2 or 3 stimuli heard as "sheet". The remaining children
with phonological impairments had great difficulty with this task.
One child indicated uncertainty about the identity of all of the stimuli and displayed
no category boundary between "seat" and "sheet". Other children identified all
seven stimuli as "seat" one hundred percent of the time. Overall, this and other
research demonstrated that a majority of children with delayed phonological skills
have difficulty with perceiving the sounds that they misarticulate. A similar
conclusion follows from a number of other studies regarding perception [4;5;6;7].
Subsequent work has demonstrated that improving children's ability to identify
such speech stimuli produces concomitant improvements in speech production abilities,
with or even without explicit articulation therapy [2;3].
The first training program based on natural speech stimuli like those used in
SAILS was reported in 1995, in the Journal of Speech and Hearing Research [3].
In this study, "sh" misarticulating children were randomly assigned to three different
groups. One group listened to and identified a variety of correct and incorrect
versions of the word "shoe". A second group identified the words "shoe" and "moo".
A third group identified the words "cat" and "Pete". This third condition was
a control condition, not expected to impact on the children's sound production
abilities.
All children received traditional articulation therapy, targeting "sh" concurrently
with the speech perception training program. Children who received the perception
training program made significantly greater gains in production accuracy relative
to the other two groups. The following figure shows one aspect of the results
from this study:
Very recently, a study published in the American Journal of Speech-Language Pathology
examined the effects of conventional production therapy using the "Cycles approach".
Such therapy was only effective for some groups of children, and SAILS was able
to identify those children which benefitted most from the production therapy.
Most importantly, ALL groups of children received the most benefit when treatment
included SAILS. Finally, treatment using SAILS not only produced greater improvements
for the children - it was also less expensive than the alternative treatment.
One important aspect of the SAILS approach is the use of children's speech sounds--
both correct utterances and real speech errors-- for training. There are several
reasons why this is important.
First, many studies have shown that children hear the acoustic cues in speech
very differently than adults do -- and this is also true for experienced speech-language
pathologists and for clinical phoneticians.
Second, adult vocal tracts are different than those of children and the speech
errors that children make are not ones that an adult can mimic effectively.
Together these factors combine to make the use of children's speech and of naturally
occurring speech errors important for teaching children the skills they require
to differentiate correct from incorrect speech sounds.
Examples of real speech errors for the /sh/ sound are provided in the three spectrograms
below. These spectrograms provide an indication of just why children so rarely
have difficulty with adult speech, even though they may have great difficulty
with their own speech and with other children's speech errors.
