Processing information in flight: Understanding the limits of cognitive capacity in the cockpit.

Hands up if you have ever experienced a mental meltdown, ‘cognitive freeze’, or intense tunnel vision in flight or in training? Most of us will recognise these phenomena happening to us at some point or other. They are intimately related to levels of workload, stress, or perhaps the surprise and startle effect. In CRM training it is often explained to us in terms of the well-known Yerkes and Dodson arousal curve – the inverted ‘U’ of arousal vs performance. 

The Yerkes Dodson Curve

That famous experiment behind Yerkes’ and Dodson’s research, which was actually the product of administering increasingly powerful electric shocks to mice, is now well over a hundred years old; around the time when aviation was still in its infancy. Since then, like the more highly charged mice, quite a few aviators of all ages and stages have probably experienced the unpleasant sensation of sliding down the backside of the curve.

What is “Capacity”?

Passing the point of ‘optimal arousal’ is more colloquially known as reaching the limits of your ‘capacity’. But what is capacity? We often use the term more generally to describe the amount of ‘thinking power’ that we have available to apply to problem-solving and decision-making.

Capacity” is used to refer to the amount of ‘thinking power’ we have available to apply to problem-solving and decision-making.

Instead of understanding the concept of the arousal curve as a line that represents the ‘before and after’ of a cognitive collapse, we should really conceive it simply as a description of the level of cognitive activity that we are capable of. Nobel Prize-winning psychologist Daniel Kahneman has described arousal in these terms, calling it ‘a reservoir of mental energy.’ In practical terms, arousal and capacity are correlated. It follows therefore that, the Yerkes-Dodson curve could also be described as your capacity level. 

Capacity and Attention

Another concept in human information processing that overlaps with this is attention. Attention is simply the way we allocate our cognitive resources. Our attention level is itself directly proportional to our level of arousal, so for the sake of understanding our capacity we can treat these too as one and the same.

Don Harris (2011) argues that attention is “in effect the amount of cognitive capacity or thinking power” that a person has available, so how we understand the functioning of our attention is important to understanding the limits of our capacity.

Most theorists agree that the way we process information is broken down into stages. We allocate our attentional resources to the different tasks of perception, memory retrieval, response selection (decision-making) and response execution. This can be seen represented by the ‘spider’s legs’ coming from the bubble of attention seen in the human information processing model below.

Human Information Processing Model

Buckets of Attention

Where the academics don’t agree however is how we expend our attention on these different stages of processing. The simplest theory explaining this process argues that all elements of our attention can be imagined as a single resource. Imagine that we have one bucket full of liquid cognitive resource which is depleted by whatever demands we put on it until it runs out. Our ability to carry out multiple tasks at the same time depends upon how much of the liquid in the bucket we allocate to each one. If we focus entirely on a single complex operation, we cannot attend to any other information hitting our senses. If we devote a lot of attention to one task, we reduce the attention we can dedicate to another task at the same time.

Capacity is a story of finite resource versus potentially unlimited demand.

Others suggest that our attention is divided into multiple resources, or a number of smaller buckets which we are able to allocate to different tasks independently. For example, we have a different bucket for perception and processing to the one which we allocate to response selection and execution. We also have a separate bucket depending on whether the kinds of inputs we are paying attention to are audio or visual, spatial or verbal. The ‘good news’ about this conceptualisation is that we can apply different cognitive resources to different tasks, but the ‘bad news’ is that our processing power is now limited by both our ability to make sense of all the data hitting our senses at the one end, and our ability to consciously process it and act on it at the other end.

Known as Multiple Resource Theory, researchers have demonstrated that while two concurrent verbal or spatial tasks do result in reduced performance, the same is not the case if one task is spatial and the other verbal for example. We can achieve separate modality (verbal/spatial, audio/visual) tasks without a significant reduction in performance in the other. Later research also found the same to be true within the visual sense. We are able to attend to one focal task, and devote attention to a second visual task in peripheral vision at the same time (although we cannot carry out two tasks at the same time that require visual focus). In other words, our visual bucket of attention can be divided into two further sections of ‘independent’ resources.

What does this tell us about managing our capacity in the real world? Well, it follows that our performance will be better if we draw upon our cognitive resources proportionately across the different buckets. By the intelligent use of all the different smaller buckets we can maintain our capacity by avoiding draining any single one. According to this understanding of how we process information, to make the most of our processing capacity we should be mindful to avoid taking on more than one auditory task at a time (offload radio communications if we are holding a conversation with the crew), and not to engage in two visually demanding tasks at once such as combining an instrument scan with the reading of an approach plate. Of course, this is where the topic of human information processing necessarily crosses over into the realms of workload management, communication, and all the other elements of CRM.

Capacity and Memory

How we interpret the information hitting our sensors forms the basis of our perceptions, and from this we create model of the world around us. We do so by constantly comparing this input data to our prior knowledge and experience of the world, and a big chunk of our attention is expended on deciding how to interpret and act upon these. Drawing on the processes, knowledge, and experiences that we hold in our long term memory and comparing it to the here and now is the cognitive process that allows us to make decisions, chose a response, and then carry out that response. 


Understanding our memory processes can also contribute to understanding the limits of our capacity. We are taught about working memory in CRM training. You’ll probably recall that working memory is very short-term, and very low capacity. There’s a ‘rule’ that the average person can retain only 7+/-2 ‘chunks’ of information (Miller, 1956) and that it can be held in working memory for a ‘half-life’ of 7 seconds (the delay after which our ability to recall that information is reduced by half).

There are a number of ways in which we can mitigate this rapid decay of items in our working memory. Research has found that acoustic items in particular can be maintained by articulating them repeatedly (sub-vocally). Chunking items works better when the chunks are strongly identified with an item in the long-term memory, where for example the four digits 1-0-1-3, can become a single chunk ‘1013’, identified in the minds of pilots everywhere with the standard altimeter setting. Whenever a memory item is meaningful in terms of a framework of prior knowledge, it is memorable. In other words your memory is strongly associative. Because of this, letters are more easily remembered than numbers, and numbers more so than a mix of letters and numbers. Working memory is also easily confused by similarity between chunks, for example adjacent number strings 5433 and 5334 are likely to be particularly difficult to recall.


Long term memory (LTM) can directly impact our capacity both in terms of the retrieval of knowledge, and the running of more automatic motor programmes. Skilled processes such as manual flying are driven by motor programmes that have been laid down in the long-term memory. Once established these programmes require relatively few cognitive resources. Creating these programmes depends upon our procedural memory, and the use of tactile and spatial methods of rehearsal provide a structure that helps to bind the neural networks we create for these skills. Hence the advantages provided by the use of simulators or touch drills in aircraft.

The other type of LTM is declarative memory and unlike procedural memory (which is hidden from view) these are memories that are consciously available to us. We store them in two different ways according to their characteristics. Episodic memory refers to the ability to recall past events or experiences and are stored as images. Semantic memory is memory for facts, rules, concepts, and problem-solving, and we store these as networks – structures to organise knowledge – by creating frameworks based on links with other meaningful knowledge.

If we already have a suitable framework for organising the information we learn established in our LTM then it is easier for us to categorise, understand, and build links and relationships with other information we already hold there. Information is also transferred to LTM more effectively when there is an existing structure there to support it and link to it. That is why when learning new topics it always helps to start with the known and move from the known to the unknown – to build upon other structures that are already in place.

It is already well understood that the building of neural networks is strengthened through repetition and rehearsal. Repeated retrieval of items from your long-term memory reinforces their pathways and structures. This explains why flying ‘armchair’ sectors from your living room is so effective. In so doing we are reducing the processing power required by recalling and retrieving the right elements from our memory stores. 

Maximising Cognitive Capacity

This may all seem rather theoretical, but it does offer some useful practical lessons to us on maximising our capacity in flight. The strength of an item in LTM depends upon the frequency and recency of its use. Training, rehearsal and revision not only allow for the strengthening of both of those elements, helping with recall, but it also fortifies the structures upon which you can hang new learning, and crucially, make inferences and analogies to other knowledge. Your mental frameworks have a crucial role to play in the creation and retention of knowledge. 

The cognitive demand required for Procedural Memory reduces progressively with practice, allowing processes and programmes to become highly automated and in some cases unconscious. Once laid down, procedural memory requires little cognitive resource, so the more that you can establish in your procedural memory, the more capacity you free up for other tasks. 

Image-based Episodic Memory is especially quick to establish in LTM and particularly enduring. Not only does a picture paint a thousand words, it also imprints this data on the brain more effectively. We often learn best from physical events and experiences that happen to us. The early and conscious reinforcing of those memory structures through debriefing and review will help cement that process.

The bottom line is that capacity – like the human information processing system itself – is a story of finite resource versus potentially unlimited demand. When we consider our capacity in the cockpit we should be conscious of the fact that we can manage both ends of this equation. The demand side is a question of workload management, but cognitive supply can also be managed. It is a function of many human factors, some of which have been described above, others of which stray into other more physiological topics such as arousal, fatigue, stress, and startle, amongst others. The aim in a perfect world? To sit ourselves on top of Yerkes and Dodson’s inverted ‘U’ every time we climb into an aircraft.


  • Harris, D. (2011) Human Performance on the Flight Deck. Farnham: Ashgate, p.24
  • Kahneman, D. (1973) Attention and Effort. Englewood Cliffs, NJ: Prentice-Hall
  • Miller, G.A. (1956) The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information. Psychological Review, 63, 81-97.
  • Wickens, C.D. (2002) Multiple Resources and Performance Prediction. Theoretical Issues in Ergonomics Science, 2, 150-77.
  • Yerkes, R.M and Dodson, J.D. (1908) The Relation of Strength of Stimulus to Rapidity of Habit Formation. Journal of Comparative Neurology and Psychology, 18, 459-82.

One thought on “Processing information in flight: Understanding the limits of cognitive capacity in the cockpit.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: