Lightning talk about lightning talks

Recently I presented a lightning talk on the topic of lightning talks.

Yeah, I know, I'm not very creative with my topic ideas.




In an extremely simplified model of memory, there are three components. Sensory register, short-term memory, and long-term memory.

Sensory registers

They're the place information is first stored as it comes in through your senses. come in two forms, visual and auditory (more commonly, iconic and echoic). Typically, auditory registers (and auditory memory in general) has greater persistence than visual, which means that if you want the user to be thinking about a given visual aid, it has better be damn well right in front of them. On the other hand, you have a bit more leeway and flexibility in terms of auditory memory - even if the user isn't paying attention at the exact moment you're speaking, they will still be able to access some mildly-decayed record of your words.

You can test this out for yourself! If you recite a series of ~8 random numbers to yourself, it is very likely you will be able to recall most of them a few seconds later. On the other hand, if you flash a series of pictures quickly, it is very hard to recall seconds later, (caveat: unless you convert them to words as you see them. words, being mostly auditory, have much stronger persistence than pictures).

Working memory

Let's talk about working memory. This is probably the most critical part for any educator to consider, because working memory is where most cognition and understanding occurs.

It's especially important in a lightning talk because you only have such limited time to teach and convey your ideas - it's easy to slip into the trap of overloading working memory by presenting too much information at a given time.

There are a lot of studies on the limits of working memory. Some say 7 items is the magic number (Miller, 1956), some say 4 is the limit (Cowan, 2001). Honestly though, I think the number isn't really useful, how do you even quantify one unit of information anyway?

So how do we keep working memory load low enough to get our ideas across?

Types of cognitive load

There are two main types of cognitive load. Intrinsic (the load naturally associated with the complexity of the novel information) and extraneous (the load associated with the presentation format).



Intrinsic load is like, the number of pieces of information. it's hard to bring down, because if you've gotta get the information across, there's no getting around that.

Extraneous load is the cognitive effort exerted when trying to understand the material presented. It's the effort caused by mode of presentation. This can be minimised by being very careful about the way in which you present. It varies a lot, and will reduce information uptake.

 Working memory 

Working memory has two main channels: visual and auditory, same as sensory registers. For now, it's just important to note that the auditory channel is system-controlled (i.e. the presenter controls the pace and rate of information presented) while the visual channel can be either user-controlled or system-controlled.

Visual memory is mostly the job of the parietal lobe and auditory memory is processed in the temporal lobe of the brain.

There are definitely some tricks to take advantage of the visual and auditory channels, though. For example, having text on screen involves both channels, because it first is processed in the visual channel and then re-encoded into the auditory channel. This is bad, because we are duplicating the same information unnecessarily across two channels and consuming working memory in both areas.

Chunking (reducing intrinsic cognitive load)

Our brains naturally is able to group pieces of related information together. The canonical example for this is the words we see in everyday life. For people who speak the language, a word is merely one piece of information. For someone unfamiliar with the language, a word could be 5 pieces of information (5 letters). Native/advanced speakers typically are even able to remember groups or combinations of words as a single item.

What does this mean for working memory? It means capacity is increased with chunking. The more times the user has seen a group of pieces of information, the more likely they are able to encode it as a single item. Chunking is one of the few ways to decrease intrinsic load - we're reducing the amount of pieces of information in order to lighten working memory load.

A great example of taking advantage of chunking is using analogies and existing knowledge to explain ideas.

"A clementine is a round fruit with a smooth glossy surface, and a juicy and sweet flavour"

"A clementine is a sweet orange"

Using existing knowledge is powerful! It does get harder the more diverse your audience gets, though.

Modality effect (reducing extraneous cognitive load)

The "modality effect" is an observed reduction in extraneous cognitive load under a certain situation:
- Information is coming in simultaneously from visual and auditory channels
- The information is system-controlled, and can be paced such that the visual and auditory channels are both engaged
- The information from each channel cannot be understood in isolation, only in conjunction with each other

The canonical example is a high school geometry problem, where the teacher must point to the angles in the diagram to demonstrate which ones are equivalent, and explain why.






There are a few caveats to this effect:
- When a user can pace their own learning, e.g. when they are looking at some diagrams, they can pace the information intake at their own discretion. The modality effect is not useful here, and in a self-paced environment purely visual stimulus often outperforms visual + auditory stimulus

- When there is information duplication across visual and auditory channels. For example, if there is a picture of a house, and the presenter says "this is a house.". This information is not distinct, and can be understood in isolation. This is called the 'redundancy effect'. and increases cognitive load rather than decreasing it. The same effect occurs when the presenter is reading text from the screen. Having to process the exact same information through multiple channels turns out to be a lot of work for working memory!

- When there are multiple visual elements, and there is no clear focus point for the viewer. If the presenter does not provide a focal point for the viewer, the visual channel is not system-controlled, and the visual stimulus then does not synergise with the auditory stimulus.

Naturally, lightning presentations are entirely system-paced. Both channels are at the discretion of the presenter. This makes them an excellent candidate for the modality effect.

Final words

Do these ideas only apply to lightning presentations? Of course not! They apply to all system-paced presentations. However, they do become very critical when time is limited, because maximising the amount of information you can get across is entirely dependent on minimising cognitive load.

In lightning presentations you also have to worry a lot less about attention span and engaging the audience! It's hard for someone to become bored after just five minutes. Interactivity is not a big concern (because the presenter has no time). Usually, the main objective is to get the idea across as fast as possible.

Further reading:
- Sweller, Ayres, Kalyuga, "Cognitive Load Theory"
- https://en.wikipedia.org/wiki/Cognitive_load