About the SA Blog Network

Talking back

Talking back

A science blog, sans blague
Talking back Home

Working Memory and The Movies Streaming In Our Heads

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

Peter Carruthers

Peter Carruthers began his career studying philosophy as an undergraduate at the University of Leeds, an outpost for Wittgenstein scholarship. Carruthers waded through the Austrian-British philosopher’s thinking for the early part of his career, getting a doctorate from Oxford and publishing books on Wittgenstein along the way.

He decided at one point to join a reading group discussing the ideas of Daniel Dennett, Jerry Fodor and Thomas Nagel, whose work deals with philosophy of mind. “It was just so much more fun,” he says.  “I just started to move sideways. Having done my Wittgenstein books, I never picked up Wittgenstein again.”

Carruthers is now in the philosophy department at the University of Maryland and his interests, like those of his adoptive mentors, focus on philosophy of mind and cognitive science. His recent work involves an exploration of  the nature of working memory. Last year, he delivered a lecture on working memory and how it evolved for the Arthur M. Sackler Colloquium at the National Academy of Sciences and then wrote a paper (PDF) on the topic for the Proceedings of the National Academy of Sciences. I asked Carruthers to explain what we know about this temporary mental storage area and whether the human version is special in any way.

What is working memory?
Working memory is the capacity to sustain in an active and conscious state a representation which can be then used in further processing or further reasoning. One example that’s often used is when someone tells you their phone number and you have to keep that number in mind while you go into another room and get a paper and a pencil.

In studying the nature of human cognition, why is working memory so important?
Differences in working memory capacity turn out to correlate with other cognitive capacities that matter. Differences in working memory correlate very strongly with fluid g or general intelligence. They correlate with reasoning ability. Many have come to think that working memory is what we are considering when we think of fluid g, though perhaps not the whole of it.

Why is this mental scratch pad so important to general intelligence?
Many sorts of routine intelligent activities we have to do, such as performing a little bit mental arithmetic, take place in working memory. Trying to focus on the road while carrying on a conversation with someone in a car also takes place in working memory because you have to hold in your mind what’s going on in the conversation while concentrating on something else. At the same time, you have to suppress or ignore certain distractor items. Being able to maintain focused attention is at the heart of fluid general intelligence

How, if at all, does working memory in humans differ from that of other primates and other animals
In its basic structure it might not differ at all. I don’t think we have enough studies to show this, but there is evidence that indicates that other animals can sustain representations. After the stimuli have gone away, they can also call up representations using visual imagery. They can manipulate those representations for long-term or medium-term planning. All the stuff we do with working memory, it looks like other animals can do that too.

So is anything special about human working memory?
There’s no good evidence about exactly what’s going on. But it may be that what’s distinctive is that we’re chronic users of working memory. We fantasize all the time. We’re using working memory at every moment of the day. When we’re not on task, we’re off task imagining something. That means you’re always playing around with alternative plans and alternative ways of doing things and so forth.

Is there any evidence that certain brain circuits unique to humans allow us  to use these mental faculties that involve imagination?
I think it’s much more likely to be motivational. There might be a subtle difference in the wiring of circuits but it won’t be a major one.  It’s more about motivation. One is motivated to call up and generate working memory as a kind of intrinsic reward to be able to use the imagination.

I’ve speculated along with a grad student in a paper I wrote recently that maybe this is the function of distinctively human pretend play. There are a lot of animals that engage in rough and tumble play. But imaginative play seems to be uniquely human. All humans do it. It comes online at about the same age in every culture. Maybe what that’s for is to get the brain in the mode of generating representations that are suggested by the environment.

There’s the famous example of pretending that a banana is a telephone. It’s roughly the same sort of shape. What you do is imagine it’s a telephone and impose a representation on it and use it accordingly. That self-generated activity can be internalized and eventually, as you get older, you don’t have to externalize your play at all. You just use your imagination.

Image Source: Cindy Phillips

Gary Stix About the Author: Gary Stix, a senior editor, commissions, writes, and edits features, news articles and Web blogs for SCIENTIFIC AMERICAN. His area of coverage is neuroscience. He also has frequently been the issue or section editor for special issues or reports on topics ranging from nanotechnology to obesity. He has worked for more than 20 years at SCIENTIFIC AMERICAN, following three years as a science journalist at IEEE Spectrum, the flagship publication for the Institute of Electrical and Electronics Engineers. He has an undergraduate degree in journalism from New York University. With his wife, Miriam Lacob, he wrote a general primer on technology called Who Gives a Gigabyte? Follow on Twitter @@gstix1.

The views expressed are those of the author and are not necessarily those of Scientific American.

Rights & Permissions

Comments 2 Comments

Add Comment
  1. 1. jtdwyer 3:26 am 03/30/2014

    Very interesting – thanks especially for the link –
    While the description of the term ‘working memory’ does not correspond exactly, as a retired information systems analyst I’d like to describe a particular ‘short-term’ memory system that might be used for initial data capture in complex database implementations. The fundamental problem being addressed is that the long-term database representation requires significant processing time to establish relations among the data and to minimize data storage requirements. Meanwhile, those necessary complex processes could interfere with reliable storage of data that must be captured in real-time.
    - The primary solution is a buffer or serial log-file that is used to most quickly record needed information about ongoing events. Everything that might be considered relevant to events is simply recorded as they happen – in simple sequential order.
    - While this log of events is quite limited in terms of information retrieval, as accessing the information relating to any specific event requires a comparative sequential search of all recorded information, it is optimized for recording speed and reliability.
    - The information recorded may be highly redundant – requiring a great deal of very limited recording space.
    - Further optimizing the recorded information to minimize storage requirements and enable more direct, associative access can best be accomplished when operation the real-time monitoring and database retrieval systems can be shut-down, or ‘put to sleep’. This allows complex programs to reprocess the captured data.
    - Redundant data bits, such as image frames containing large areas of blues sky might be compressed and encoded to indicate a field of blue pixels. Silent sound samples or sounds considered to be extraneous may be effectively eliminated.
    - Objects considered to be of limited relevance to the event, such a crowd of people on the sidewalk, a sleeping dog, cars on the street, etc., may be replaced by links to representative reference images. Even conversations, weather conditions and any such extraneous information might be generalized to minimize the storage and reconstruction time necessary to recall events that are considered important.
    - Indexes relating a new event to previous events and classes of events may be constructed and/or updated to allow associative, contextual retrieval of events.
    - Standard elements, such as the sleeping dog image mentioned previously, might be modified to better represent all the events in which they’re referenced.
    - As the event log-file is processed, or when processing is completed, it can be erased for reuse when the event-capture system is again reactivated, or when the system ‘wakes up’.
    - Similar processes might occur within humans and other animals, evidenced by functional and capacity restrictions that distinguish ‘short-term’ and ‘long-term’ memory processing and operational consequences resulting from failure to perform ‘sleep-state’ memory migration processes. The design of information systems, after all, are the product of human analytical processes…

    Link to this
  2. 2. jtdwyer 6:12 am 03/30/2014

    BTW, similar such processes may explain some of the characteristics of dream recollections as experiences are parsed into snippets and common segments are retrieved for modification and re-linkage. They may also explain why our recollections of events change over time, as common elements are altered to better reflect more recent experiences…

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article