On Task
By David Badre
“Cognitive control means we can find a way to make real what are conceivable despite our constraints.
Likes
It’s a nice combination of academic and casual writing - credible and challenges itself often
The concept of cognitive control is illustriously defined, and the author makes it easy to grasp
Good use of metaphors to make tricky concepts easier to understand
Ends at exactly 300 pages
Dislikes
Book starts with “none of this may be true”, but i prefer that over a book that claims itself to be scientific but ends up as pseudoscience
The brain stuff is a little hard to keep up with (e.g. which brain parts do what) but I feel it’s necessary
Chapters are pretty long, and text is crammed on page making it difficult to read
Synopsis
I picked up On Task from Stanford Book Store while out doing some research in Boston. It was a nice sunny day and myself and a colleague were out touring MIT and Stanford while preparing for some research. I noticed this book in the Cog Psy section and picked it up. As a designer I feel it’s good to know more about why people do what they do, so we can better design products for them.
The book covers the concept of cognitive control. Cognitive control allows us to accomplish out goals despite the world’s constraints. We do this through a combination of sequencing tasks, accessing working memory via gates in our brain, understanding and learning rule systems, and applying task schemas in the real world to do a thing. The book also covers topics like inhibition (why we stop doing things), transference (how cognitive control exercises might improve performance), the effects of aging, and dives into the nitty gritty of how different areas of our brain contribute to the cognitive control system.
The book presents an interesting hypothesis, and provides a large amount of evidence in support. As a designer I could only understand so much of it, but I feel like I got what I wanted from it, which is a deeper look into how people accomplish their goals, and why.
Biggest Learnings
Cognitive control (or executive function) is our brains function of linking knowledge with action
Transfer is the process by which we take our knowledge from one task and bring it to another (e.g. breakout to pong)
We compose and generate schemas to complete tasks based on memories, even things we have never done before. For schemas to be done in the future we build them using future episodic thought.
Working memory may be accessed through (input/output) gates which control which info we let into and out of working memory. By presenting context first, we can limit the amount of information people need to store in their working memory to complete a task.
We must translate our rules/goals into gating policies, and the policy we choose will greatly effect our performance/efficiency.
We have goals and sub goals to complete those goals. After enough repetition, executing these goals can become automated.
Neural activity for similar tasks is called family resemblance
Capture error is when we make an error based on learned sequencing of tasks, while trying to complete another task
Our brains like hierarchy, and they impose hierarchical rules to decide which gates to open. These rules can have more than four orders of hierarchy.
People are bad at multitasking
Task-switch Costs are costs to performance incurred due to switching tasks.
Tasks that are mutually exclusive have a reduced task-switch cost. Or, the more similar two tasks are the greater likelihood there is for a bottleneck or performance loss.
Inhibition is the process by which actions are stopped from happening, and there are likely various reasons why this happens.
Value is subjective, and our expectations for the value we receive tend to fluctuate or balance out for given stimuli with high variance
Dopamine, motivation (or lackthereof) is a driver of action (go) and inaction (no go)
Go neurons and Nogo neurons strengthen their synapses based of outcomes, this is how learned responses form
Our brain may perform cost-benefit analyses to decide which task to do
We make decisions about what task to do based on not only the value we stand to gain but the costs in mental work required to gain that value.
The act of making decisions makes it more difficult to make future decisions
Ego depletion / sapping willpower may lead to lower task performance
Circular reasoning is a logical fallacy in which the premise of an argument is used to generate a conclusion that is then used to support the initial premise.
Stimulus-outcome (e.g. smell to salivation) vs. Response-outcome (bite to tasty cake)
When retrieving info, we weigh the likelihood of success against the effort required to recover an item and the risk of taking another, easier approach.
When remembering, the determining factor is not the reliability of storage, but the balance between the benefits of success and the costs of ensuring that success.
Control effects memories in 2 ways. 1) by leveraging existing knowledge and foregrounding certain info as a cue, we use controlled retrieval to get more useful info. 2) We also engage in control post-retrieval to ensure we act on memories that are most useful to our goals.
Being able to appropriately value memory doesn’t just ensure us that we remember useful things; it shapes our understanding of the works around us.
Cognitive control may be hereditary, and diminishes with aging
Cognitive control is multi-faceted between inhibition, updating, and working memory
Transfer effects are cognitive exercises that transfer to different tasks. Near transfer is when one test improves performance on another test that is only superficially different. The goal with cognitive control training is far transfer, which is improvement on tests that are unlike the trained test or in a different domain. Far transfer hasn’t been accomplished or proven yet, but there’s a lot of evidence for near transfer.
Cognitive control means we can find a way to make real what are conceivable despite our constraints.