Monday 14 December 2015

The ecological approach to sporting performance

Last week I took part in a Google Hangout with Mark Upton and Al Smith, who run the sports coaching blog 'My Fastest Mile' and who are generally all about getting sports coaching and training to be more ecological and dynamical. The recorded hangout is here - I assume it's good, I can't watch myself on video without cringing :) I've fleshed out some of these ideas below; please comment below or on Twitter if there's anything you want more on.

We talked perception, affordances and a little about what makes the ecological approach different from more cognitive approaches. There's a lot of detail sitting under the discussion; see, well, the rest of the blog to get a sense of where I'm coming from! I'm not a sports scientist, and so my particular research programme isn't specifically about applying the ecological approach to sports. However, I do study some sports related activity (specifically throwing; see my latest paper, blogged here). 

The main thing that gets applied to sports from the ecological approach is the concept of affordances (which is the topic of that paper). Affordances are cool, obviously, but the thing that makes an approach 'ecological' is a focus on information; how things like affordances are perceived. I wanted to briefly sketch how I see ecological psychology feeding into sports, which I think it's a great idea for everyone involved.

A disclaimer: I am also not a coach. I'm just a scientist. I like to let people know that I know this, because I've found that whenever I interact with practictioners of any kind (coaches, but also occupational therapists, physios etc) there's a real resistance to listening to people like me. I mean, what do we know? We aren't in the trenches, working with the athletes or patients, and our wonderful ideas might simply not apply to the messy real world. I've also chatted to people who felt worried I would be judging their messy actual practice, the one they've put together over years of experiencing the actual needs of the people they help. 

I understand this concern entirely. Let me say, for whatever it's worth, that I am not trying to waltz in, figure out what you're doing wrong and save you with my wonderful theory. I always see my role as just 'the scientist in the conversation'. I'm going to listen to what you actually want to know, and I'm going to see if there's anything I know from what I do that can feed into that process. It will either work or it won't, and that's ok. I get to think about ways to make my science work in an interesting applied high stakes context; I hope you might get some insights into how learning works that help you make sense of what you see in the people you work with. 

That all said, what is it I think I have to offer? Broadly speaking, I think I bring a theory that reflects a genuinely new way to think about perception, action, learning and coordinated activity. I think this theory sits on top of a lot of strong, relevant evidence (it's not just me, I'm just one with the microphone right now!) and I think this strongly supported theory is a really productive way to conceptualise what's going on with a learner as you try to help them find their way through the process to an ambitious goal. I also think it has many implications for sports (for some ideas, see Fajen, Riley & Turvey, 2008)

The approach is called the ecological approach to perception-action. The ecological approach was first developed by James J Gibson over the course of a 40 year research career and three books. The most relevant and accessible is The Ecological Approach to Visual Perception (1979), which has recently been re-released in print and ebook form. It's kind of brilliant.

Gibson develops two basic ideas to enable him to talk about how skilled action can be supported by the perception of our environments, as opposed to the operation of internal mental models. 
  • Affordances: this is the idea you have probably encountered. Affordances are opportunities for action. They are dispositions of the environment; that cup is graspable by me, for example. Gibson proposed that properties like this can be perceived by organisms, and that by perceiving them we can coordinate and control skilled activity without needing complex internal models. I can work to grasp that cup, not because I figure out that I can grasp it, but because I literally perceive that it's an option and can move to implement ('effect') that option.
  • Information: this is the idea you probably haven't encountered. Gibson knew that things like affordances could only work if they could be perceived. Traditional sensation based theories of perception don't let you perceive things like affordances, and he also had a lot of evidence that perception wasn't based on sensations anyway (we don't experience anything like what happens on the retina, for example). He therefore developed a new account of the kind of information perception uses, and there has been 30 years of research since the book confirming that this information exists and that we use it to perceive all kinds of things. 
The set up therefore is:
  1. At any given moment, we find ourselves in environments that offer some affordances and not others. I'm in a coffee shop now full of places I can walk and sit but nowhere I can play football. The current environment also provides only some information, specifically information about the opportunities for action in that environment. We call our current environment the task space.
  2. The most appropriate way to formally describe a task space is in terms of dynamics. Things aren't random, and dynamics is the mathematical description of how and why events unfold over time the way they do. Dynamics is the appropriate level because it uses the units of length, time and mass, so you can talk about motions and the forces causing those motions (i.e. a complete account of the local task space). When you study the environment to figure out what it offers an organism, we talk about studying the task dynamics
  3. Task dynamics also create information. Gibsonian information consists of patterns of motion in energy arrays, e.g. light for vision. These patterns are what happens when energy interacts with task dynamics. These patterns are kinematic, that is, best described only using units of length and time. They are therefore not identical to the dynamics that caused them, but we somehow need to use these patterns as information about those dynamics. 
  4. Information is not identical to the dynamics that caused it, but it can specify those dynamics. That means that it can map 1:1 - one kinematic pattern maps to one dynamical pattern. Because of this, an organism can use the information as if it were the dynamics and everything works out
  5. This information is not simply an image or view on the world. Information can provide access to the complete dynamics from all kinds of view points. Two people who have learned the information for a dynamic literally perceive the same task, which provides a stable perceptual basis for creating coordinated activity with no need to 'shared mental models'. 
  6. Organisms use this information to interact with task demands (dynamics), and so behaviour is shaped by how we perceive our environments. This means behaviour is task-specific; there are no core skills, only skills-in-contexts. A great outfielder doesn't throw the same way to cover the same distance because they are executing a remembered action ('throwing'); they are interacting with the same information they are practised in using to organise behaviour ('throwing-in-context'). 
The practical upshot of all of this is that, as Gibson proposed, we can perceive our environments in the kind of rich detail that enables us to produce skilled, functional and task-relevant behaviour in real-time without any kind of complex internal 'mental gymnastics'. Learning is still part of the game, but learning is primarily perceptual; we are learning to detect and use information to coordinate and control our behaviour. 

I'm a coach - why would I care?
This is a great question. The basic answer is simply that Gibson's account of how we come to produce skilled behaviour proposes an entirely different chain of events for the learner. Learning isn't about developing interesting mental models; it's about become perceptually attuned to the complex opportunities for action a task can offer. Intervening in the learning process (i.e. coaching!) needs to know how motor learning really works in order to tailor its interventions correctly. (Other factors matter; motivation, etc all come into play. Think of the ecological approach as 'the better way to think about the motor control layer of how skilled actions are acquired and executed'.)

One implication is about how to make learning transfer from training to the game. How can you structure practice so that what you learn there makes you better on the field? There's two basic approaches - practice elements of skills to work on technique (e.g. drilling on kicking) or practice in a game context (e.g. practice kicking during a normal game). Transfer of learning is a massive topic in motor control and one with few, if any solutions right now. We know transfer when we see it, but we struggle to predict it and we struggle to explain it when it happens. The basic suggestion is that learning will transfer if the two contexts overlap in some meaningful way, but that's as precise as anyone has gotten. 

Using the ecological approach, I have suggested that this overlap must be task dynamical, because this will make the contexts share information, and behaviours are caused by our real time interaction with information. The world pulls behaviour from us by providing information that we learn to use, rather than our brains pushing behaviour out of a programme or mental model. I have a proof-of-concept paper on this (Snapp-Childs, Wilson & Bingham, 2015) where I predicted and explained transfer of learning as a function of what information the two contexts have in common. 

So a coach, when developing a training programme, needs a task-dynamical style analysis of the skill-in-game-context that they want to improve, and should provide a training environment that contains the same affordances and therefore information as in the game. Training should be complex enough to contain task relevant information, but should also make it so the learner can focus on learning that information. You don't simply add noise or variability or complexity to a training scenario - you work to add task-relevant complexity, and task dynamics is how you figure out what is relevant.

Pulling this off in practice is obviously hard and I am not a coach; designing real training scenarios is not something I'm trained to do. But taking a task dynamical, ecological approach shifts the coach's attention, away from memory and notions of core skills and onto emphasising that skills exist in contexts. Contexts in turn are dynamic, and create information, and must be perceived in order to affect behaviour. Training shouldn't be about creating core skills which can then be deployed on demand, it should be about creating stable, flexible, perceptual expertise with a context. 

Task dynamics is obviously hard to do for complex tasks; it was hard enough to do for a standing throw to a target (Wilson et al, in press). But that maths heavy analysis came out of a verbal description of the task of throwing, just one framed in dynamical terms. Throwing, dynamically, is the creation of a projectile motion that satisfies some demand (maximising distance, intercepting a target). Everything I do with respect to throwing should be done with respect to those dynamics, and without doing much if any maths I have a first cut, task relevant way to constrain my training. 

Fajen, B. R., Riley, M. A., & Turvey, M. T. (2009). Information, affordances, and the control of action in sport. International Journal of Sport Psychology, 40(1), 79. Download

Snapp-Childs, W., Wilson, A. D. & Bingham, G. P. (in press). Transfer of learning between unimanual and bimanual rhythmic movement coordination: Transfer is a function of the task dynamic. Experimental Brain Research. Download

Wilson, A. D., Weightman, A., Bingham, G. P., & Zhu, Q. (in press). Using task dynamics to quantify the affordances of throwing for long distance and accuracy. Journal of Experimental Psychology: Human Perception and Performance.  Download (pre-publication version), Supplemental Material

No comments:

Post a Comment