[This is an oldie-but-goodie (OGB) post that I originally posted as a guest blogger at the IM-HOME blog on Feb 2, 2012]
No human investigation can be called real science if it cannot be demonstrated mathematically
Leonardo da Vinci, Treatise on Painting (1651)
Progress in science depends on new techniques, new discoveries and new ideas, probably in that order Sydney Brenner (1980)
At the core of the IM intervention technology is a precise
measurement system. To users and clinicians the IM measurement system
is transparent. Yet, without the valid and precise measurement system,
IM would not work.
In my “Brain or neural efficiency: Is it quickness or timing?” post, I advanced the hypothesis that the effectiveness of Interactive Metronome may be due to IM operating on a fundamental dimension of brain or neural efficiency, which intelligence scholars also relate to general intelligence (g). I have also suggested that this mechanism improves control of attention and may allow individuals to “quiet a busy mind”and invoke “on-demand focus.”
As an applied intelligence test developer (click here),
I have been intrigued by the underlying precise millisecond-based
measurement system which is the heart of IM technology. IM technology
would not work if the underlying measurement system could not reliably
measure differences in synchronized metronome tapping between
individuals and changes within the same individual over repeated
sessions.
Wanting to know how precise the underlying IM measurement system is,
I extracted the average millisecond scores from an unpublished 2003
Interactive Metronome document that reported average times for different
age groups. The sample consisted of the initial IM Long Form
Assessment performance of 1,583 clinical and normal subjects ranging in
age from 6 thru 60+. It is important to note that the sample was not a
nationally representative normal sample and was comprised of more
clinical subjects receiving IM therapy. Nevertheless, I wondered if
this less-than-optimal set of data might demonstrate a pattern of
increasingly shorter response times as individuals became older. Why
did I want to examine this?
Developmental increase in proficiency on tests and measures of human
abilities is considered one form of evidence that a test or measurement
system is reliably and validly measuring an important human ability.
In the case of intelligence, valid measures of cognitive abilities show developmental growth curves
where the youngest subjects obtain the lowest raw scores and the
average raw scores gradually increase with increasing age. They
eventually level out and then start a decline as old age sets in. Below
are growth curves from seven cognitive ability scores from the Woodcock-Johnson Battery—III,
a test battery of which I am a co-author. The important observation to
note is that, despite the specific cognitive ability measure, all
curves show low scores for the younger ages followed by acceleration of
growth to a certain point. Each curve then plateaus at a certain age
range, after which age-related cognitive decline is noted, but at
different rates for different abilities. These curves are presented in
the WJ III Technical Manual (McGrew & Woodcock, 1991) as a form of developmental validity evidence—which provides one piece of evidence that the WJ III tests are valid measures of different and important human intellectual abilities.
The first thing the reader should note are the individual data points (the dots). The points show some random “bouncing around” which we measurement folks call sampling error. The critical point is that they follow a systematic trend that can be estimated by fitting a mathematical curve to the data points. This was the same procedure used to develop the WJ III cognitive curves in the first figure. In the second figure, the IM timing curve is demarcated in red. We who develop test norms and study human ability growth curves generate these smoothed growth curves as they are the best estimate of the real reality of the data if extremely large number of individuals had been tested at each age (there would be much less bounce).
One does not need to be a rocket scientist to interpret the smoothed IM growth curve. Individuals at the youngest ages, on the average, show the largest millisecond discrepancy from the IM reference tone. Then, with increasing age, the average IM target-to-response for individuals decreases systematically as children age. At approximately 25 years of age the curve “bottoms out,” and then as individuals get older, IM millisecond timing scores increase (or get less accurate). The systematic nature of this curve is amazing, considering it is based on a less-than-optimal sample for determining what constitutes average.
If the reader is having a hard time relating the IM timing curve to the
WJ III cognitive ability curves, I have taken the liberty of simply
rotating and flipping the IM timing accuracy growth curve in the figure
below. Vioila (aka, walla—“there it is”)! The curve has the
same general shape as the WJ III cognitive ability growth curves! The
reason for the difference between the WJ III growth curves and the first
IM timing growth curve is that the meaning of high and low scores are
reversed—higher IM times mean lower skilled performance while lower
scores on the WJ III battery are associated with lower performance (and
vice versa).
Readers who are parents may have seen similar growth curves during well-child visits with the family doctor. Below are growth charts for weight and length for male children from birth to 36 years. Although covering a much smaller age span than the WJ III cognitive and IM timing curves above, the shape of the curves is identical for the comparable age ranges (gradually increasing with age). The middle dark line in each set (labeled 50 for 50th percentile) is conceptually identical to the above single curve plots. These physical measurement curves show the systematic and developmental nature of physical growth.
Why am I so excited about the IM timing growth curve? Because it demonstrates, similar to the physical and intelligence growth curves, that the underlying measurement unit used as the core of IM therapy is measuring a human ability that follows a similar and expected developmental pattern. Such curves are believed to be due, depending on the specific ability, to the influence of education and experiences as well as genetically-driven biological maturation of the central nervous system (CNS). The IM timing curve is one form of evidence that the IM measurement system is measuring a fundamental human capacity. This is extremely exciting! It is one more piece of evidence that the IM core measurement technology is measuring and working on a core critical human ability. Coupled with other validity evidence previously discussed here and elsewhere, this additional piece of scientific evidence has convinced me that the IM measurement and intervention system is most likely measuring a fundamental aspect of the development of the central nervous system (e.g., neural efficiency). The cognitive abilities I have suggested fall under the broad umbrella term of executive functions, and more specifically controlled attention (focus) and working memory.
A caveat before I close. The smoothed IM timing curve should not be used by IM providers to evaluate how typical, normal, or close-to-average a person is on their initial IM Long Form Assessment. The mixed nature of the sample (normal and clinical subjects; more of the later) argues against such use. Also, the curve only represents the average at each age and calculating and plotting the typical variability around the curve would also be necessary. I deliberately left out the variability data curves so as not to encourage misuse of the information.
However, IM providers can evaluate their client’s performance by using the official IM Indicator Table. A copy is reproduced below. This table can be used to determine whether a client’s performance is in the “ballpark” for their age. Providers simply locate the clients age in the row at the top then go down that column to find the millisecond score or range that includes their specific IM Long Form Assessment timing score. The verbal description associated with each level (extremely deficient to exceptional) can be used to make quality of performance statements reflecting where an individual is at the time of the initial assessment. The scores and labels should not be used for diagnostic purposes. Instead, they can be used to describe, in approximate ball park terms, where an individual is at the time of the assessment when compared to others of the same age and to make comparisons about that same client’s performance over time.
Age
|
6
|
7 to 8
|
9 to 10
|
11 to 12
|
13 to 15
|
16+
|
| Extreme Deficiency |
280+
|
270+
|
260+
|
240+
|
215+
|
200+
|
| Severe Deficiency |
175-279
|
170-269
|
160-259
|
155-239
|
150-214
|
147-199
|
| Below Average |
120-174
|
90-169
|
80-159
|
75-154
|
72-149
|
70-146
|
| Average |
90-119
|
65-89
|
55-79
|
45-74
|
43-71
|
41-69
|
| Above Average |
56-89
|
45-64
|
38-54
|
36-44
|
33-42
|
30-40
|
| Exceptional |
40-55
|
32-44
|
28-37
|
26-35
|
23-32
|
22-29
|
| Superior |
Below 40
|
Below 32
|
Below 28
|
Below 26
|
Below 23
|
Below 22
|
In summary, I have traversed a number of empirical domains in my journey to understand IM. The finding of such powerful and clear developmental evidence for the underlying IM measurement system is one of the final dots I connected which convinced me of the promise of IM. The IM program is founded on a valid scientific measurement system of an important human cognitive ability (or constellation of related abilities).
Ken Gilhooly (
