Unsupervised Thinking
a podcast about neuroscience, artificial intelligence and science more broadly

Tuesday, May 28, 2019

Episode 45: How Working Memory Works

Working memory is the ability to keep something in mind several seconds after it's gone. Neurons don't tend to keep firing when their input is removed, so how does the brain hold on to information when it's out of sight? Scientists have been probing this question for decades. On this episode, we talk about how working memory is studied and the traditional view of how it works, which includes elevated persistent firing rates in neurons in the prefrontal cortex. The traditional view, however, is being challenged in many ways at the moment. As evidence of that we read a "dueling" paper on the topic, which argues for a view that incorporates bursts of firing, oscillations, and synaptic changes. In addition to covering the experimental evidence for different views, we also talk about the many computational models of working memory that have been developed over the years. Throughout we talk about energy efficiency, the difference between maintenance and manipulation, and the effects of putting scientific disagreements in writing. We also admit to not reading *any* primary sources.

We read:
Persistent Spiking Activity Underlies Working Memory
Working Memory: Delay Activity, Yes! Persistent Activity? Maybe Not
Working models of working memory

And mentioned previous episodes:
Bayesian Modeling in Psychology 
Brain Freezing and Cooling
Neural Oscillations

And some other work:
Ring attractor for heading direction in flies
Neuronal circuits underlying persistent representations despite time varying activity

To listen to (or download) this episode, (right) click here or use the player below

As always, our jazzy theme music "Quirky Dog" is courtesy of Kevin MacLeod (incompetech.com)


  1. Thanks for the timely podcast. It was just what I was looking for.
    I am a student of the brain who studies its operation as an avocation - not associated with any college or business.
    My current quest is to learn what is known about the last 500 mSec of my existence. What is the underlying biological functions that allow us all to experience reality.

    I have downloaded the papers you referenced and am reading them now but as you described your thoughts on Working Memory it occurred to me that the working memory contents of that monkey's brain must have contained much more than the light flash. It must have contained the entire awareness of the animal.

    That 'working memory' must have contained the monkey's sensation of heat/cold, hearing, other images in his field of view, that itch under his left paw and on and on. It must also contain his current thoughts as they arise and are pondered on. In short, the working memory should actually be called 'The Present'.
    I wonder if that has occurred to others?

    I have a simulated neuron based in the Izhikevich model, but compiled running now to learn from. If your interested you can download it (Windows only) here:

  2. Yea! There is for sure more information stored or present in the monkey's brain as it's holding the flash location in mind. That is frequently what people mean when they say "noise" or talk about how working memory needs to be robust to noise. Of course it's not really noise if it is useful info, but from the point of view of the particular task being studied it's irrelevant. So it's mostly viewed as a nuisance/thing to overcome in the context of these highly controlled and unnatural experiments