Learning from Cognitive Load Theory

So I will admit that I have been much behind in my writing about my learning here, but in my defense, it's because I have become absolutely lost in the fascinating literature of medical education. I tend to be the kind of person who reads as much as I possibly can and must fully understand it as a whole before I am able to synthesize and summarize for other people. Harking back to my previous post, I guess that makes me an Analyst more than a Wholist... that 'Deep before Broad' concept...

This is a bit counter-intuitive given my most recent learning topic has been Cognitive Load Theory and teaching styles that have been perfected to allow medical students to learn most effectively. Cognitive Load Theory (summarized very effectively by van Merrienboer & Sweller),  has been developed to reflect the concept that human brains are only capable of maintaining a finite number of elements in their working memory (7-9 to be exact). Clearly we would not be able to have evolved to our current level of sophistication if humans were not able to become more efficient: thus, learning. By encoding these elements together into a 'schema' in our long term memory, we are able to reduce those previous 7 or 9 elements into a single element which can be recalled at will. Interestingly, there has been no limit identified for holding in your working memory information that has been recalled from long-term memory. This essentially means that our brains are limitless, provided we learn each new concept in discrete packages and incorporate it into our long-term memory and appropriate schemas.

Applying Cognitive Load Theory to lecture and teaching design, we can easily see that it would be fruitless to have an hour long lecture that simply spouts facts at you that greatly exceeds your brain's working memory. It's interesting, because this is actually how I felt at the very beginning of medical school. When we started learning anatomy, I felt like I didn't even have the language to understand what the words were (superior or posterior? myocardium or pericardium? artery or vein? It was all the same to me.) It took hearing the words over and over, reviewing my notes at the end of the day, and seeing it in person in the anatomy lab for my brain to be able to develop a schema. It took a week, but eventually I could get through a lecture and actually understand what was being said rather than frantically copying out notes.

Now that I've read more about it, I wonder if there isn't a more efficient or effective way to teach anatomy. With regards to working memory, there are several ways the load can be affected: the actual information/learning elements being presented (intrinsic load), the manner in which it is taught (extrinsic load), and by how much learning actually happens - i.e. how hard your brain is working to deal with, create schemas, and process the new information/intrinsic load (germane load). Using the anatomy lecture analogy, the intrinsic load (terminology and content) was huge; and, given that it was new, the germane load was almost too much for me to consciously absorb or understand anything until some of the information had been encoded into a schema (specifically, the vocabulary). Neither of these factors are particularly adjustable, but the extrinsic load (way it was presented) could in theory have been adjusted to reduce the total stress to working memory.

There are many strategies for reducing extrinsic load. Some examples are:

1. Goal Free Principle: Replaces conventional goal-directed tasks with goal-free tasks. For example, rather than choosing the 'best diagnosis' (which requires reasoning and judgement), students are asked to provide a list of as many diagnoses as possible.
2. Worked Example Principle: This relies on the fact that is easier to study and learn from a fully solved problem in front of you before trying to work through a problem on your own.
3. Completion Principle: Breaks down a complex task into manageable parts (i.e. somebody gets you started, and then you complete the task).
4. Modality Principle: Splitting the intrinsic information from a single modality into multiple; for example, providing verbal instructions with a visual diagram rather than written instruction. 

More than reducing extrinsic load though, I've learned that it's about optimization of working memory. For example, if the concepts are simple and intrinsic load is low, your working memory is likely capable of a high extrinsic load as you have more room to work with. However, that would be highly inefficient. Ideally, we would want the intrinsic load to be as high as we can manage with extrinsic load minimized. As well, ultimately, material must eventually be presented in its fully complexity to be fully understood. From this arise the principles of teaching from the ground up with simple concepts first and gradually increasing complexity. A great example of this is allowing students to practice first with paper-based cases then proceeding to high fidelity simulation environments prior to being with a real patient.

There are also strategies for optimizing germane load, or, how we process the information. There have been several studies that suggest our brains learn more effectively when they are challenged or forced to work. In that sense, intrinsic load is intentionally made more complex to increase the germane load and optimize the amount of working memory being used in a giving learning situation. There are some specific examples of how teaching strategies can embrace this concept:

1. Variability of Task Situations: This is the idea that multiple examples (varied demographics, comorbidities, etc) of a particular clinical scenario, like solving an acid-base equation allows us to determine what are the 'key points' more easily by seeing them across the various presentations. van Merrienboer describes this perfectly: "[it] encourages learners to construct cognitive schemas because it increases the probability that similar features can be identified and that relevant features can be distinguished from irrelevant ones." 
2. Contextual Interference: Also known as mixed teaching vs block teaching. This principle is based on the idea that is easier to identify what are the pathognomonic findings of a particular diagnosis when they are contrasted and interspersed with other diagnoses. As Hatala et al. demonstrated in their ECG teaching study, this allows us to see why it is a left bundle branch block vs a right bundle branch block for example when see them one after another, rather than just memorizing what a left bundle branch looks like by seeing 5 in a row.

Another concept is that of distributed learning, which, while not directly mentioned in the context of Cognitive Load, is similar to a larger scale application of contextual interference. As Raman et al. demonstrated in their paper, long-term retention of nutritional information was increased by 4 1-hour sessions distributed over four weeks rather than in a single 4 hour block. From my own personal experience, I think this also has to do with a student's inability to concentrate for that length of time without taking mental breaks. However, it is true that when the topic is constantly switching, it does make it easier to pay attention (similar to the mixed teaching mentioned above.) 

Upon further reflection, my experience of excessive cognitive load while learning anatomy was likely inevitable. Perhaps the stress could have been lessened by mixing anatomy lessons in with other topics, but, as I recall this was also done - we had histology lectures, embryology lectures, and radiology as well as anatomy lab time mixed in. What is often forgotten is that learners are not all coming into a given learning situation with the same baseline amount of knowledge. What felt like a steep learning curve for me, having no anatomical vocabulary, would have been straightforward for a student coming from a background in physiology or human kinetics. While it would have been nice to have had a longer adjustment time for myself, this would have been too slow for the other learners in the class, and would have perhaps have had the opposite effect on their learning.

This is the final point made by van Merrienboer and Sweller in their paper; that of Expertise Reversal. As they explain "This effect is an interaction between several basic cognitive load effects and level of expertise. The effect is demonstrated when instructional methods that work well for novice learners have no effects or even adverse effects when learners acquire more knowledge." This concept outlines how an attempt to minimize extrinsic load (ie by using worked examples) may become redundant as a students' intrinsic load capacity increases by evolving schemas. 

My overall conclusions are that:

•  University of Toronto designed a pretty darn good medical school curriculum, though there is always room for improvement. 
• Cognitive Load Theory can teach us a lot about effective teaching strategies and should be taken into account when designing not only curricula but individual teaching sessions
• Going forward, I am certainly going to be incorporating some of these strategies into teaching medical students. My first hands-on try at this will be on Thursday; let's see how it goes!

~LG