It's early morning, the sun is barely rising, and your only desire is a cup of coffee to kickstart your day. As you head towards the kitchen, a fresh plan brewing in your mind, your phone buzzes. A new work email! Could it be important? You decide to check. Turns out, it's just spam, so you quickly delete it and continue to the kitchen. But then, you pause, your mind suddenly drawing a blank.

Wait, what was I doing?

We've all experienced it: in the midst of an activity, moving with purpose, when suddenly we hit a reset, our plans vanishing from our minds.

The brain function responsible for tracking our plans and helping us stay focused is our working memory. I've dedicated much of my career to studying this field of cognition, exploring its limits, and understanding how it correlates with attention and learning. 

The term "working memory" is somewhat misleading: it's less about memory and more about working on, or attending to, a task. Randall Engle suggests that it is largely synonymous with executive attention. The scientist often credited with coining this term, Alan Baddeley, has remarked that if given the chance to rename it, he would have chosen "working attention" instead.

The neural mechanisms of working memory and attention are remarkably similar: they are manifested through temporary patterns of electrical activity. Computer simulations suggest this happens through local feedback loops, known as recurrent activation. This coding method is quick but easily disrupted. Our capacities for working memory and attention are limited. In contrast, long-term memory is encoded through more durable changes in the brain: the synapses forming the physical connections between neurons become stronger.

How could one illustrate working memory? Alan Baddely suggested a “model” consisting of three connected boxes: a phonological loop, a visuo-spatial sketchpad, and a central executive. This model has gained widespread recognition, though linking this cartoon-like model specifically to the brain's functions has been challenging. The "central executive" component has been particularly elusive - is it a homunculus or ragbag? 

Committing an episode to long-term memory is akin to drawing a picture in the sand with a stick. The grooves may erode over time, but they are relatively permanent. Interestingly, when we recall a long-term memory, it often changes slightly, similar to retracing a path in the sand.

The temporary electrical activity of working memory and attention resembles the ripples caused by a stone thrown into a lake: the impact creates a pattern of ripples, which can be quickly disrupted by another object. This analogy reflects the pattern of energy flow - waves don’t travel; they are just particles moving up and down. It is the energy that travels. However, unlike ripples, working memory does not inevitably fade away. 

Computer simulations provide a physiologically more realistic model for working memory. They simulate neural networks with recurrent activity and suggest that working memory activity can be modeled as an attractor network. This is a dynamic pattern that stabilizes for a while, like a whirlpool in a river that persists - until suddenly the water's flow changes and it disappears. This transience might be what occurs when you walk into a room and forget why you’re there.

The term “working memory” was actually used more than a decade before Alan Baddeley, by Karl Pribram, a psychiatrist and neurosurgeon known for his research on the prefrontal cortex. In 1960, he wrote:

“... we should like to speak of the memory we use for the execution of our plans as a kind of quick access, ‘working memory’.”

A memory for plans, like, you know, making coffee.

‍

References:

Miller, George; Galanter, Eugene; Pribram, Karl (1960). Plans and the Structure of Behavior. New York: Holt, Rinehart and Winston.