Long-term memory is the brain’s lasting store of information: everything you keep beyond the few seconds that short-term memory holds, from facts and words to life events, skills and habits. It can keep material for hours or for a lifetime, and no study has ever found its capacity limit. It comes in two broad kinds: memories you can consciously recall (explicit) and memories that show up only in what you do (implicit).

The types of memory, and where long-term memory fits

The classic multi-store model of Richard Atkinson and Richard Shiffrin (1968) describes three stages. Most textbooks still use it as the starting point, even though later research has refined each stage.

  • Sensory memory: a very brief trace of what you have just seen or heard, a fraction of a second for vision and a little longer for sound.
  • Short-term and working memory: a handful of items held for seconds while you use them, such as a number you are about to dial. Our page on working memory covers its limits and its strong link to reasoning.
  • Long-term memory: the durable store, with no known capacity limit. It is not one system but several, as the next section shows.

The types of long-term memory

The division most researchers use today, summarised by the neuroscientist Larry Squire, splits long-term memory by whether you can consciously bring the memory to mind.

  • Explicit (declarative) memory is what you can recall and put into words. It has two parts. Episodic memory holds personal events tied to a time and place: your first day at school, what you had for breakfast. Semantic memory holds facts and knowledge cut loose from where you learned them: that Paris is the capital of France, what a word means.
  • Implicit (non-declarative) memory shows itself in performance rather than recall. Procedural memory holds skills and habits, such as typing, riding a bike or driving a familiar route. Priming is the way earlier exposure makes something easier to recognise later, and conditioning is learned association, such as feeling anxious at the sound of a dentist’s drill.

What patient H.M. revealed

Much of this map comes from one man, known in the research literature as H.M. In 1953, at the age of 27, he had parts of both medial temporal lobes, where the hippocampus sits, removed to control severe epilepsy. Afterwards he could no longer form new long-term memories of facts or events, and that memory loss came with no general intellectual loss. His digit span (the short-term store) was intact, and over ten trials he learned to draw a shape while watching his hand in a mirror, keeping the skill across three days, yet had no recollection of having done the task before. He was studied for five decades until his death in 2008.

H.M. showed three things at once: short-term and long-term memory are separate systems; the medial temporal lobes are needed to form new explicit memories but not to learn skills; and memory and intelligence can come apart. (His older memories were not all spared: later testing found gaps reaching back about 11 years before the operation.)

How memories get into long-term storage

  • Encoding: attention comes first. What you never attended to is rarely stored, which is why our page on cognitive load matters for learning.
  • Consolidation: new memories are fragile at first and are stabilised over hours to years, with the hippocampus gradually handing them over to the cortex. A large review by Björn Rasch and Jan Born (2013) concludes that sleep, especially slow-wave sleep, plays an active part in this consolidation.
  • Retrieval: memories are reconstructed from cues rather than replayed like a recording, which is why they can change or be wrong in detail. Our page on photographic memory explains why truly picture-perfect recall has never been demonstrated.

How much can it hold?

More than anyone has managed to fill in an experiment. In a 2008 study, people viewed pictures of 2,500 objects over 5.5 hours. Shown pairs afterwards, they picked the one they had seen 92% of the time when the other was a different kind of object, 88% when it was another object of the same kind, and 87% when it was the same object in a different state or pose. Long-term memory, in other words, stores detail on a large scale.

How much people actually retain is a different question. Thomas Landauer (1986) estimated the information adults keep from experience at around a billion bits, roughly 125 megabytes, far below what the brain could in principle hold. Popular claims that the brain stores a fixed number of petabytes are extrapolations, not measurements.

The forgetting curve

Hermann Ebbinghaus measured forgetting on himself in the 1880s by learning lists of nonsense syllables and timing how long it took to relearn them later. The time saved on relearning (his “savings”) fell fast and then levelled off:

  • 20 minutes: 58% savings
  • 1 hour: 44%
  • 1 day: 34%
  • 6 days: 25%
  • 31 days: 21%

A 2015 replication at the University of Amsterdam, in which one person spent 70 hours learning and relearning lists, produced a very similar curve. Note what savings measures: the effort saved on relearning, not the share of a list you can still recall. Something “forgotten” still leaves a trace that makes relearning faster, so the often-repeated line that we “forget 70% in a day” misreads the data.

How to remember things for longer

  • Space it out. A 2006 meta-analysis of 839 comparisons in 317 experiments found that spaced study beats cramming, and that the longer you need to remember something, the longer the best gap between study sessions.
  • Test yourself. In a 2006 study, students who recalled a passage from memory remembered more of it a week later than students who reread it, even though rereading felt more effective.
  • Sleep on it. Sleep after learning supports consolidation, as the review above sets out.
  • Make it meaningful or vivid. Linking new material to what you already know, or to images and places as in the memory palace technique, gives retrieval more cues to work with.

Is long-term memory related to IQ?

Yes, in two ways. First, much of what IQ tests call crystallized intelligence, such as vocabulary and general knowledge, is semantic long-term memory: it measures what you have learned and can retrieve. Second, the ability to learn and recall new material is linked to fluid reasoning too. In a 2010 study by Nash Unsworth, recall and recognition from long-term memory and working memory formed separate but correlated factors, all strongly related to fluid intelligence and less so to crystallized intelligence, and both working memory and the memory factors as a whole explained unique parts of fluid intelligence.

Memory is still not intelligence. H.M. could not form a single new long-term memory of an event, yet showed no general intellectual loss. A good memory helps on many tasks, but reasoning with what you know is a separate ability, as our guide to the g factor explains.

Trust and scope notes

This page summarises published research for general information, with figures taken from the studies listed below. It is not medical advice: memory problems that are new, worsening or worrying deserve a conversation with a doctor, and no online test can assess them. The IQ Revealed test is an online reasoning test of 40 original items for adults aged 18 and over; it is not a clinical assessment and it does not measure memory. Our guide to how IQ tests work explains what an online test can and cannot tell you.

IQ Revealed is not affiliated with, endorsed by or connected to the researchers, universities, journals or publishers named on this page. Names are used only to identify the research being described.