Research School Network: The Relationship Between Working Memory and Long-term Memory Our long-term memory has potentially infinite capacity – so how can we use it to support working memory?

Working memory has a very short duration. Unless information is actively maintained in working memory, for example by rehearsing or refreshing, it is likely to be lost within 30 seconds. As we rehearse or refresh that information, we are engaging in processes that can help transfer information to long-term memory.

In theory, we can hold information in long-term memory indefinitely. However, information in long-term memory can decay over time, or suffer from interference. Interference is when similar memories can become confused, for example, you are trying to remember what you had for lunch last Sunday, but you have memories of many different Sunday lunches from the last few weeks and months, and these interfere with your ability to single out the specific instance you want to remember. So, long-term memory is not perfect but it does allow us potentially to store information over very long periods of time.

Can we use long-term memory to ​‘bypass’ working memory?

With regards to capacity, long-term memory is also potentially limitless, unlike the very limited capacity of working memory. However, in most circumstances, information from long-term memory that is required to complete the task at hand still has to be brought into working memory. Exceptions to this mainly involve routine motor activities that have become automatic (e.g., tying your shoelaces), and priming. Nonetheless, for most classroom activities, you cannot simply bypass the limited capacity of working memory by tapping straight into long-term memory.

How can we use long-term memory to reduce working memory demands?

However, information that is securely stored in long-term memory is more familiar, which in turn helps with accessing and processing that information in working memory. Also, we create schemas, or webs of knowledge, within long-term memory for related information. This means that we can then ​‘chunk’ that information, by combining several separate pieces of related information into one ​‘chunk’, which in turn frees up some of the capacity of working memory. Bringing one chunk of related information into working memory from long-term memory reduces demands compared to remembering lots of separate items of new information. For example, a schema or network of information in longterm memory on WWII will place a lighter load on working memory resources than when we first encounter this information and each fact is effectively a single, isolated ​‘item’. This is why tasks that require lots of new pieces of information, are not familiar, or do not tap into any schemas in long-term memory, are likely to place higher demands on our working memory

This is an extract from Working Memory: Research into Practice. Download the guide and the poster below.