Story
How Memories Form
Last updated
In one pass Remembering yesterday's dinner with a friend involves material scattered across the brain: what you saw, what was said, the taste of the food, each handled by different areas of cortex.
Educational content, not medical advice — consult a clinician.
Story path
Chapter 1
The hippocampus binds an experience together
Remembering yesterday's dinner with a friend involves material scattered across the brain: what you saw, what was said, the taste of the food, each handled by different areas of cortex. So what binds them into one memory?
The answer lies on the inner side of each temporal lobe: a seahorse-shaped structure, named the hippocampus for that shape, together with adjacent cortex. Studies of amnesic patients and of monkeys show this system is essential for establishing long-term memory for facts and events; through extensive two-way connections with the neocortex, it binds the distributed storage sites into a whole memory (Squire 1991).
The key line: in the standard model that Squire set out in 1991, its role is only temporary. As time passes after learning, memory stored in the neocortex gradually becomes independent of it. This is a model, and not every researcher thinks it applies to every kind of memory. Taking it on its own terms, the better image is not a warehouse but a temporary index: it records where each part of an experience is kept, until those parts are linked firmly enough on their own.
The answer lies on the inner side of each temporal lobe: a seahorse-shaped structure, named the hippocampus for that shape, together with adjacent cortex. Studies of amnesic patients and of monkeys show this system is essential for establishing long-term memory for facts and events; through extensive two-way connections with the neocortex, it binds the distributed storage sites into a whole memory (Squire 1991).
The key line: in the standard model that Squire set out in 1991, its role is only temporary. As time passes after learning, memory stored in the neocortex gradually becomes independent of it. This is a model, and not every researcher thinks it applies to every kind of memory. Taking it on its own terms, the better image is not a warehouse but a temporary index: it records where each part of an experience is kept, until those parts are linked firmly enough on their own.
Chapter 2
Synapses strengthen with use
If memories live in the connections between neurons, then remembering something should correspond to some change at the level of those connections. In 1973, Bliss and Lømo saw one such change in anesthetized rabbits.
They stimulated a bundle of fibers into the hippocampus with a burst of high-frequency pulses, then tested with single pulses: in 15 of 18 rabbits, hippocampal neurons' response to the same pulse was stronger, lasting from 30 minutes to 10 hours (Bliss 1973). This is long-term potentiation (LTP): after a pathway has been used intensely, it transmits more effectively, and stays that way.
Know its limits: this is an electrophysiological phenomenon measured at one synapse in anesthetized animals, not a sighting of a memory itself. Its importance is that it shows one concrete way a connection between neurons can be changed by experience and stay changed. How neurons pass signals to each other is covered in the article on the nervous system.
They stimulated a bundle of fibers into the hippocampus with a burst of high-frequency pulses, then tested with single pulses: in 15 of 18 rabbits, hippocampal neurons' response to the same pulse was stronger, lasting from 30 minutes to 10 hours (Bliss 1973). This is long-term potentiation (LTP): after a pathway has been used intensely, it transmits more effectively, and stays that way.
Know its limits: this is an electrophysiological phenomenon measured at one synapse in anesthetized animals, not a sighting of a memory itself. Its importance is that it shows one concrete way a connection between neurons can be changed by experience and stay changed. How neurons pass signals to each other is covered in the article on the nervous system.
Chapter 3
From hippocampus to cortex, with sleep's help
In Squire's 1991 standard model, the hippocampus's role is temporary. So how does a memory go from depending on the hippocampus to standing on its own in the cortex? The process is called system consolidation.
The 2013 review by Rasch and Born puts sleep at the center of it: the waking brain is better suited to encoding new memories, the sleeping brain to consolidating them. In slow-wave sleep (the deepest stretch of sleep, with the slowest brain waves), recently encoded memories are reactivated and transformed for integration into long-term storage; the sleep that follows may stabilize the transformed memories. In this framework, called active system consolidation, sleep is not merely protecting memories from interference but also actively moving them. It is an explanatory framework, and not every link in it has been measured directly in people.
How a night of sleep is staged, and what slow-wave and REM sleep each do, is covered in more detail in the article on sleep architecture. The point to take from here: sleeping after learning is part of how a memory forms, not a rest from it.
The 2013 review by Rasch and Born puts sleep at the center of it: the waking brain is better suited to encoding new memories, the sleeping brain to consolidating them. In slow-wave sleep (the deepest stretch of sleep, with the slowest brain waves), recently encoded memories are reactivated and transformed for integration into long-term storage; the sleep that follows may stabilize the transformed memories. In this framework, called active system consolidation, sleep is not merely protecting memories from interference but also actively moving them. It is an explanatory framework, and not every link in it has been measured directly in people.
How a night of sleep is staged, and what slow-wave and REM sleep each do, is covered in more detail in the article on sleep architecture. The point to take from here: sleeping after learning is part of how a memory forms, not a rest from it.
Chapter 4
Why spaced review lasts longer
Reading a chapter three times the night before an exam, or once on each of three days: same total time. Which lasts longer?
Cepeda 2006 pooled 839 assessments from 317 experiments in 184 articles, looking specifically at spacing: bunching study together versus spreading it out. The result is a very practical rule: the best gap between study sessions depends on how long you need to remember — the longer you need to remember, the longer the best gap.
The review reports the rule, not its cause. One explanation that fits system consolidation is that a gap before restudying usually contains consolidation time and sleep. Whatever the cause, the practical meaning is clear: preparing for an exam next week and trying to keep a language for life simply call for different review rhythms.
Cepeda 2006 pooled 839 assessments from 317 experiments in 184 articles, looking specifically at spacing: bunching study together versus spreading it out. The result is a very practical rule: the best gap between study sessions depends on how long you need to remember — the longer you need to remember, the longer the best gap.
The review reports the rule, not its cause. One explanation that fits system consolidation is that a gap before restudying usually contains consolidation time and sleep. Whatever the cause, the practical meaning is clear: preparing for an exam next week and trying to keep a language for life simply call for different review rhythms.
Chapter 5
Testing yourself beats rereading
The most common way to review is to reread. Roediger and Karpicke 2006 had students read a passage and then either reread it or close the book and recall it in writing (with no feedback), the same number of times.
The result comes in two parts: tested after 5 minutes, the rereaders remembered a little more; tested after 2 days and after 1 week, those who had recalled remembered substantially more. More striking, repeated rereading made students more confident they would remember — confidence that pointed the opposite way from their score a week later.
The study measured the effect, not the mechanism; one reading that fits system consolidation is that a memory is not stored just by being read, and memories that are used and reactivated hold better. In practice, treat closing the book and saying it aloud, doing problems, or explaining it to someone as the review itself, not a check after reviewing.
This article is about how normal memory forms. How memory changes with age, and when to worry, is covered in the article on cognitive aging.
The result comes in two parts: tested after 5 minutes, the rereaders remembered a little more; tested after 2 days and after 1 week, those who had recalled remembered substantially more. More striking, repeated rereading made students more confident they would remember — confidence that pointed the opposite way from their score a week later.
The study measured the effect, not the mechanism; one reading that fits system consolidation is that a memory is not stored just by being read, and memories that are used and reactivated hold better. In practice, treat closing the book and saying it aloud, doing problems, or explaining it to someone as the review itself, not a check after reviewing.
This article is about how normal memory forms. How memory changes with age, and when to worry, is covered in the article on cognitive aging.
References · 5
- Squire, L. R., & Zola-Morgan, S. (1991). The medial temporal lobe memory system. Science, 253(5026), 1380-1386. From human amnesia and a monkey model: the hippocampus with adjacent entorhinal, perirhinal and parahippocampal cortex is essential for establishing long-term declarative memory (facts and events), binding the distributed neocortical storage sites that represent a whole memory. Its role is temporary: with time after learning, neocortical memory becomes independent of the medial temporal lobe. 10.1126/science.1896849
- Bliss, T. V., & Lomo, T. (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. Journal of Physiology, 232(2), 331-356. Original report of long-term potentiation: after high-frequency conditioning trains to the perforant path, the dentate granule-cell population response was potentiated for 30 min to 10 hr in 15 of 18 urethane-anaesthetised rabbits. An animal electrophysiology finding, not a memory measurement. 10.1113/jphysiol.1973.sp010273
- Rasch, B., & Born, J. (2013). About sleep's role in memory. Physiological Reviews, 93(2), 681-766. 10.1152/physrev.00032.2012
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380. Meta-analysis of 839 assessments in 317 experiments from 184 articles on spacing and lag. The inter-study interval and the retention interval act jointly: the interval giving maximal retention grows as the retention interval grows. Reports the pattern, not its cause. 10.1037/0033-2909.132.3.354
- Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255. Two experiments: students studied prose passages, then took one or three free-recall tests without feedback or restudied the same number of times. At 5 min, restudying gave better recall; at 2 days and 1 week, prior testing gave substantially greater retention, although restudying raised students' confidence. 10.1111/j.1467-9280.2006.01693.x