Learning, arousal and dancing mice

Posted by | December 17, 2013 | Post | No Comments

John Helmer gets excited about the Yerkes–Dodson law and its implications for workplace learning in the age of 70:20:10.

Science has been aware for over a century that the ability to learn varies with arousal levels (quiet at the back there, we’re using the word in its technical sense here). But the full implications of this effect have only relatively recently become clear.

The foundation of this area of study is the Yerkes-Dodson law, formulated in 1908, as a result of experiments on mice. Regrettably, electric shocks were involved. The scientific justification for wiring up these somewhat macabrely-named ‘dancing mice’ was to test the relationship between the strength of a stimulus and the rapidity of learning. Would the mice learn faster to run into the ‘safe’ white box and avoid the shock-administering black box if the shock were more intense?

The results of this experiment led to more than a century of controversy and have something to say to us still about learning in the era of 70:20:10.

The Yerkes-Dodson law, in its original form, states that performance improves with raised arousal – but only up to a point. After a certain level, performance of complex tasks declines, while performance of simple tasks continues to improve as the stimulus grows stronger.

Robert Yerkes was a behaviourist and went on to become a prominent eugenicist and (arguably) racist, making him a difficult figure for us in the 21st Century. But this was not the reason for his and Dodson’s work falling into disrepute for most of the Twentieth Century. Despite their experiments being revisited and validated by more modern methods, during the 1950s it was simplified and misrepresented by D.O. Hebb and others, and the so-called Hebbian version of the law that made its way into the textbooks (see diagram) became widely vilified in the field of cognitive psychology.

Within the last few decades, however, developments in neuroscience – and in particular greater knowledge of the effects of stress hormones (glucocorticoids or GC) on human cognition – coupled with the high profile given to the issue of post-traumatic stress in war veterans and victims of disasters such as 9/11, has led to a resurgence of interest in what the unfortunate dancing mice were really trying to tell us (see this paper: The Temporal Dynamics Model Of Emotional Memory Processing: A Synthesis On The Neurobiological Basis Of Stress-Induced Amnesia, Flashbulb And Traumatic Memories, And The Yerkes-Dodson Law. David M.Diamond – Adam M.Campbell – Collin R.Park – Joshua Halonen – Phillip R.Zoladz – Neural Plasticity – 2007)

 

Is stress a learning-killer?

The crucial distinction missed or ignored by scientists of the 1950s was that made in the study between simple and complex tasks, and how these are differently affected by high-stress conditions.

 

Original Yerkes-Dodson Law

OriginalYerkesDodson1

 

Hebbian version of the Yerkes Dodson Law (this version leaves out that hyperarousal does not adversely impact simple tasks). This version is the most common version and often incorrectly cited in text books.

HebbianYerkesDodson1

What now seemed to be the meaning of Yerkes-Dobson was that strong states of anxiety did not inhibit the formation of long-term memory altogether, they just made the brain behave in a more selective and focused way – producing the type of ‘flashbulb memory’ associated with traumatic events.

Under high stress conditions, people can perform simple tasks with greater efficiency – and they can also form long-term memories of extraordinary vividness. But they are impaired in performing more complex tasks, and also experience memory dropouts in the period following the traumatic event. Jobs that require careful decision-making and multi-tasking with divided attention are done less well. Similarly, more complex forms of learning are impaired.

The subject is itself complex. Although far more is now known than ever before about how the brain works to form memory, still the detail of how various parts of the brain interact in doing this (particularly the Hippocampus, frontal lobes and Amygdala), and the neurobiological processes involved, are matters of ongoing research.

There is, it might be speculated, a tipping point in arousal – a point at which it becomes stress and ceases to help with certain types of learning. But exactly where this borderline might be, and how it might be recognized and defined, have not been mapped out in the literature, so far as I am aware (although I’d be pleased to hear from anyone more knowledgeable on the subject who has better information!)

 
Learning gets closer to work
So what are the implications of this research for workplace learning? Given the emergent nature of the field there are limits to the inferences that can be drawn from it, of course. However it clearly has important relevance to many of the changes that are currently sweeping through organisational learning; changes that we have documented in our white paper The New Learning Organisation and which are given empirical support by Towards Maturity’s latest benchmark report.

Previously, the assumption was that training must involve some degree of removal from the workplace. This is no longer the case. Organisations have been shutting down their bricks-and-mortar training centres for some years now. At the same time, the amount of learning that is delivered electronically directly into the workplace has gradually increased until it has now begun to approach parity with that delivered by humans.

More recently, we have seen the growth of mobile devices as a delivery channel for learning, spreading its access far more widely among those not ‘nailed’ to their desks. Notionally, employees can now access learning wherever they happen to be.

Not all of this ‘learning’ is learning of course. Much is ‘mere’ information, knowledge management, or company comms. Courses are honed down into learning nuggets, to be accessed as and when needed. We talk a lot about just-in-time learning and about performance support. We employ mantras such as ‘knowledge embedded within workflow’.

All of this represents a huge change in the context of use for learning content. With the workplace fast becoming the default location for its delivery, we might well ask, with Yerkes-Dodson in mind, whether we should perhaps design learning a little differently. In particular we might ask whether arousal levels are likely to be different in the workplace than they are in the training room.

And the answer to that question that will come back from almost anyone who has ever earned an honest shilling for their labour is that modern workplaces are high-stress environments.



Learning in the place of work – or stressing?
According to the latest research published by the CIPD, ‘in autumn 2013 the proportion of employees reporting excessive pressure at work every day or once or twice a week is 41%’.

You don’t have to be a soldier on active service, in the police or the fire service, or an oil rig roustabout on the North Sea in a force ten, to experience high levels of stress at work. Those who work in politics, the media or advertising accept stress (not just pressure but stress: there is a qualitative difference) as part of the territory. If you work in a business that is failing, or being taken over, or for an organisation that is having its funding cut or squeezed … if you feel your job is at risk … if you feel
yourself to be a more honest and conscientious person than those to whom you report … if you do work to which a great deal of secrecy, or financial risk is attached …  I could go on all day!

We enjoy watching amateur chefs on TV experiencing the full heat of a professional kitchen, because it reminds many of us of occasions on which we were thrown into the fire as raw employees; situations in which our mettle was really tested. Remember that time when you stayed up all night to get a pitch/proposal/new-product-launch out against an impossible deadline? Or dealt with the demands of an impossible boss without getting a conviction for ABH? Or dealt with an impossible employee, for that matter, without ending up in a tribunal?

Of course you do. These are formative experiences in many working lives; rites of passage for many industries. The flash-bulb memories formed by such experiences are among the most enduring lessons we ever learn at work – and part of the 70% called, rather drily, experiential learning, within the 70:20:10 formulation.

But let’s be clear about exactly what type of learning we are talking about. These are often traumatic experiences. The lessons you learn are not all good. There is a whole different area of learning that simply can’t take place under such conditions; the type that involves, for instance, calm reflection, balancing of multiple and perhaps conflicting principles, and attention to detail.

 
Hot and cool contexts of use
In deliberate contrast to the workplace, training rooms are by and large low-stress environments – designedly so, given that the prevailing wisdom in the post-war, Hebbian period when the ‘traditional’ model of training evolved, was that high levels of arousal were inimical to learning, full stop.

Yerkes-Dodson shows that a certain level of arousal is useful for learning, and probably essential. It also shows that simple types of learning can happen in high stress situations, but that more complex types cannot. A learner journey might involve the need for both types of learning, and thus for both ‘hot’ and ‘cool’ environments to learn in.

Understanding the context of use is increasingly becoming a factor in user interface design, due to the influence of mobile, and will slowly work its way through into learning design as well. But further than this, it seems to follow that being able to monitor and control levels of arousal in the learning situation (and in the individual learner, given that response to stimuli varies among individuals) could be a determining factor in the success of learning interventions.

This is no more and no less than a really good instructor does in the classroom: sensing the interest level of the class, turning the voltage up or down – but making this scalable, personalisable and adaptive with the use of technology calls for something more than the intuition of individual good teachers; it calls for an equal partnership of art and science in designing experiences and systems.

Making this effective in the workplace will mean being clear about the limitations of many of our instructional models, however. Yerkes-Dodson tells us that using any but the most primal instructional approach will be useless in the full heat of a high-stress workplace crisis. Support tools that give fast, clear information and communication links are what is needed in that situation … or failing that, a version of Siri that can give a useful response to the words ‘Beam me up, Scotty’.