Fontaine, Lee & Swinnen (1997): Experiment 2

|Review| The influence of the Acquisition of a New Coordination Pattern on the Intrinsic Patterns

 

If you’ve landed here before reading Experiment 1, I’d suggest you check it out so that the rest of this clobber makes sense.

In Experiment 1 Fontaine and authors were interested in testing the influence of intrinsic patterns on the acquisition of a new coordination pattern. That is, do 0° and 180° affect the learning of a new pattern? Is the “stronger” attractor (0°) a greater antagonist for the learning of a new pattern? In this case, Science said no. It turns out it was harder to learn a pattern closer to 180° than 0°.

In Experiment 2 the question is flipped. Wild, I know. Does learning a new pattern of movement affect the intrinsic patterns (0° and 180°)?

Motivation

It may come as no surprise that the motivation for this experiment also came from Zanone and Kelso (1992, 1994). In these studies they found:

Learning 90° negatively affected the accuracy and stability of the anti-phase (180°) pattern. As skill in 90° increased, the stability and accuracy of 180° waned.

In a similar study, Lee, Swinnen, and Verschueren (1995) could not fully replicate Zanone and Kelso’s results. In partial support, they found that the destabilisation effect was present on the first day of practice. This effect had vanished by the second day of practice and remained elusive for the rest of the study.

The present study was designed to further examine the conflict between the Zanone and Kelso and the Lee et al. findings. (Fontaine et al., 1997, p 5)

Putting it to the test

Fontaine et al. extended the practice period to 6 days, including over a whopping 4,000 individual bimanual coordination cycles. This was a sensible way to tackle the discrepancy between the number of trials used for each study.

The basic experimental setup was the same as Experiment 1. The difference being that instead of two groups learning alternate patterns, there is just one group learning one pattern (90°). Test trials measuring 0°, 90° and 180° were taken before, during and after practice.

Results and Discussion (for all you results nerds, click here)

In a nutshell, the results are a five-year old’s first experience with finger paints; a bit of a mess. There is partial support (in a very round-about way) for Zanone and Kelso’s prediction. The support comes in the form of a twofold change (though one is only temporary!) in the intrinsic patterns, but (and this is a considerable BUT) they are not the changes predicted.

If we look back at the earlier paraphrasing of Zanone and Kelso’s work, they predict a detrimental change in 180° as 90° is learned. This detrimental effect of learning a novel pattern is present immediately after practising the novel pattern. What goes against the prediction is that this effect is present in both 0° and 180°. Not only that, this effect is temporary. It simply vanishes the very next day. Over the duration of the study, there is a gradual change in the accuracy of the intrinsic patterns. 0° starts out as the more accurately produced intrinsic pattern, but by session 4 it is replaced by 180°. This shift in accuracy is maintained throughout the rest of the study.

The Attractor Shoe Doesn’t Seem to Fit

I see two problems (bear with me through some attractor talk). One: We should see a much stronger perturbation in 180° than 0°. This is because 180° is the weaker attractor. That doesn’t happen. Both intrinsic patterns are affected pretty much equally. Two: The model predicts this temporary change to remain to some extent in future sessions. It vanishes entirely after just one day.

The accuracy switch is odd. Again, the model would expect the opposite; 0° (as the stronger attractor) should be less affected by the learning of a novel pattern (90°). It isn’t. Over the course of the study 180° (as the weaker attractor) becomes more accurate and 0° less so. Furthermore, these changes only occur in regards to accuracy. There is no temporary or permanent change in the stability of the intrinsic patterns.

We must conclude (like Lee et al, 2015) that these results are part of a body of literature that directly oppose some of the core predictions generated from the HKB model.

Attractor Talk

If you’ve made it this far and you’re not familiar with the CRM literature, you’re probably thinking, ‘what the hell is an attractor?’ I hear you.  The short answer is, it’s something someone made up in order to explain some CRM data.  The model of attractors (also known as the HKB model) work for some of the CRM literature, but like here, it doesn’t always seem to hold up. That being said, I should probably designate a post to elaborating on the whole attractor shebang. Stay tuned for that one!

Results for the Nerds

Just like Experiment 1, the two key variables we are interested in are accuracy (measured with absolute constant error) and stability (measured by standard deviation)

• Accuracy: Before and After each session

On average, the performance of the intrinsic patterns prior to practice (M = 7.9°) was more accurate than post-practice (M = 10.1°). This suggests a temporary decline in performance immediately after the practice of the novel pattern. Temporary being the key word, as this decline was not permanent, with performance returning to pre-practice levels the following day.

• Accuracy: The overall pattern

In the early stages of learning 90° we see the typical pattern of 0° being produced with higher accuracy than 180°. As the acquisition of 90° begins to take hold there is a progressive shift in the accuracy of the intrinsic patterns. 180° becomes more accurate and 0° less so. By the final practice session (6) this difference is at its peak and remains so during the retention measures taken four weeks later.

• Stability

Learning a new pattern (90°) had no effect on the stability of the intrinsic patterns. There were no temporary deteriorative effects (like there were for accuracy). Neither were there any long term effects (positive or negative) on the intrinsic patterns. There was a significant difference (p <.05) between the average stability of the intrinsic patterns. With 0° (M = 9.7) being more stable than 180° (M = 11.3°). However, this difference was not present at session 5 and onwards. At retention, there were no differences in stability between the intrinsic patterns.

• Learning and retention of 90° in both variables

We can confidently say that ability to perform 90° was learned and retained. As far as accuracy is concerned 90° outperformed 0° but not 180°. It was produced with a high degree of accuracy, even outperforming 0° (but not 180°). Stability wise, 90° was not as stable as either of the intrinsic patterns, but the stability reached at the final practice session (6) was retained.

References

Fontaine, R. J., Lee, T. D., & Swinnen, S. P. (1997). Learning a new bimanual coordination pattern: reciprocal influences of intrinsic and to-be-learned patterns. Canadian Journal of Experimental Psychology = Revue Canadienne de Psychologie Expérimentale, 51(1), 1–9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9206321

Lee, T.D., Swinnen, S.P., & Vershueren, S. (1995). Relative phase alterations during bimanual skill acquisition. Journal of Motor Behavior, 27, 263-274.

Zanone, P. G., & Kelso, J.A.S. (1992). Evolution of Behavioral Attractors with Learning: Nonequilibrium Phase Transitions. Journal of Experimental Psychology: Human Perception and Performance, 18(2), 403–421. http://doi.org/10.1037/0096-1523.18.2.403

Zanone, P.G., & Kelso, J.A.S. (1994). The coordination dynamics of learning: Theoretical structure and experimental agenda. In S.P. Swinnen, H. Heuer, J. Massion, & P. Casaer (Eds.), Interlimb coordination: Neural, dynamical, and cognitive constraints (pp. 461-490). San Diego, CA: Academic Press.