A retrospective study project in Mjölby
The Sensomotoric Centre in Mjölby, Sweden has since 1987 used reflex- and visual stimulation programs and since 1990 also specific auditory stimulation (ADT) programs with a total of some eight hundred students (Sohlman, 2000, pg. 16).
The auditory stimulation in this (and in the other three reported programs) are based on assessment procedures comprising hearing tests (hearing thresholds, auditory laterality and some form of dichotic listening). The students were tested prior to and post intervention.
Over a period of 3-18 months the students listen daily for 10-15 minutes to specially composed and recorded music tapes, where the volumes via equalizer in 1/3 octave bands are manipulated (lowered or boosted) to partly compensate for the variation between the measured hearing thresholds and the optimum curve as suggested by Gulick (1971) and by Tomatis (1991). In all right handed and in the majority of left handed students, the sounds to the right ear are boosted the most.
(The music has been specially made for this purpose by the Danish composer Bent-Peder Holbech and covers the frequency range 100 Hz to 16000 Hz).
The students’ hearings are controlled at regular intervals, and new tapes based on the follow-up assessments are recorded and handed out.
From a file with 127 cases completed during Jan. 1997 - April 2000 fourteen files (m: 13; f: 1) were randomly drawn. The small sample was chosen due to expected large effect size based on earlier pilot studies. All were right handed with a mean age of 10yr10m (9;1-13:7). One student had dropped out before finishing the stimulation program. All students had normal hearing according to standard hearing tests (20 dB screenings). All students were referred from school or from parents based on tests for language impairments, delayed reading and/or spelling. The mean stimulation period for the reported study was 29 weeks (10 - 65). The students had been listening at home and were supervised by their parents.
An age-matched control group (N=24) mean age 10yr10m (8;3-12;9) of good readers (above the mean on teachers assessments) were tested on dichotic listening (Hughdahl DLCV-108 NF).
1) The total variation (sum) in dB at eleven frequencies between the measured hearing thresholds for both ears and the optimum hearing curve was calculated for each individual and correlated with the number of errors by dichotic listening in the non-forced condition (Hughdal DLCV-108 NF). The correlation was moderately negative (-.49).
This may indicate that initially a slight to moderate variation between the actual hearing and the optimum hearing in some individuals is more damaging to auditory acuity than a more profound variation approaching a small hearing loss or hypersensitivity.
2) Files from six of the thirteen students had enough data to deal with questions 2, 3 and 4.
These six students were all right-handed males with mean age of 10yrs01m (9;10-10;04).
The mean variation from the optimum curve for the six hearing curves (R + L) before stimulation was 205.00 (SD=54.16). After stimulation the mean variation from the optimum curve was found to be 122.50 (SD=39.44); d (effect size) > .80. (According to Cohen (1988) a d above .80 is a large effect size).
The mean error rate by DL-NF before the stimulation period was for these six students 33.33 (SD=13.05). After the stimulation period the mean error rate was 14.00 (SD= 9.24)); d>.80.
The mean error rate by DL-NF for the 24 age matched controls (good readers) was 13.50 (SD=6.13). Thus the difference in mean error rate between the controls and the research sample after intervention was not significant (fig. 1).
For the thirteen students who completed the stimulation period the mean variation from the optimum curve was reduced from 220.38 (SD=75.77) to 143.46 (SD=77.12); d>.80.
After 19 weeks of stimulation one of these students had no alterations in hearing curve at the right ear (variation from the optimum curve before and after was 230, which appeared to be the largest variation in the sample). Variation in his left ear improved from 210 to 155 (reduction in variation between curves.)
It may be essential to note that in several other clinical trials also the left ear has shown the most rapid improvements in sensitivity later followed by improvements of sensitivity in the right ear. It may be related to the known better myelination of the neuronal fibers in the right hemisphere and to the known earlier maturation of the right hemisphere (Korpilahti 1996).
The remaining twelve students had a mean variation from the optimum curve (right ear only) of 87.92 (SD=15.61) before the stimulation period and of 57.08 (SD=18.31) after stimulation; d>.80.
Thus 92.3 per cent had positive effects from stimulation of hearing during the training period with a large effect size. t test for dependent means (repeated measures design): t(11)= -12.210; p<.01, one tailed.
Therefore it can be concluded that specific auditory stimulation has an effect on hearing sensitivity. (Generally sensitivity is reduced in the low frequency range (< 1000 Hz) and increased in the high frequency range (< 1000 Hz)).
3) For the six students the reduction in total variation (R + L) between the actual hearing curves and the optimum hearing curve correlated with reduction in errors by DL-NF at r=.19.
This is a minor correlation, but contrary to the suggestion that the variation between the actual hearing curves and the optimum curve is negatively correlated with errors at DL-NF and thus with hearing sensitivity.
Looking at the right ears only in these six students it was found that the mean variation between the actual and the optimum hearing curve was reduced from 88.33 (SD=15.72) to 57.50 (SD=20.56). d>.80.
The correlation between the reduction in variation between the actual hearing curves for the right ears only and the optimum curve and the reduction in errors by DL-NF was found to be .68.
Therefore it can be concluded that reduction in variation between the actual hearing curve and the optimum hearing curve for the right ear alone (altered sensitivity) after specific auditory stimulation can lead to improved auditory acuity.
4) For the six students with a complete set of filed data the stimulation periods varied from 21 weeks to 65 weeks. The reduction in error rate by DL-NF correlated with length of stimulation period (r= .86) - fig. 2.
Surprisingly enough fig. 1 shows that reduction in errors by different individuals happens at a very similar rate during the stimulation period (apx. .5 pct/week). Due to the small number of students this indication, of course, must be dealt with caution.
Fig. 3 shows 2 x 2 audiograms from students before and after auditory stimulation. These two students had the longest stimulation periods and the largest reduction in error rates by DL-NF.
Student AB: The variation between the actual hearing curve for the right ear and the optimum curve decreased from 65 to 35. Reduction in error rate by DL-NF: From 58% to 31%.
Student CD: The variation between the actual hearing curve for the right ear and the optimum curve decreased from 80 to 30. Reduction in error rate by DL-NF: From 42% to 11%.
Discussion of the Mjölby study
CAPD is generally not assessed by simple means such as audiometry, binaural audiometry and dichotic listening.
The reported results from the clinical work in Mjölby indicate that assessment procedures utilizing these simple tools can provide valuable information about a language impaired child’s auditory difficulties and at the same time provide the necessary information for a remedial technique where listening to specially composed and specially recorded (manipulated) music is an essential part.