Table of Contents

Why Memory Matters

The Three Parts Underlying All Your Memories
Encoding: Putting memories into the brain
Intention to learn: Does it matter how much you want to remember?
Depth of Processing: Why how you process information determines
how much you’ll remember later
Transfer-appropriate processing: The trick to acing your exams
State-dependence: How your physical/mental states drive your
memory
Storage: Keeping memories in the brain
Trace decay: Do your memories fade with time?
Interference: Does learning new things block your old memories?
Retrieval: Accessing Memories in Your Brain
Spreading activation: Gain quick access to your memory
Retrieval failure: What to do if you get stuck in an exam
Practice tests: The most powerful technique for boosting memory
Summary of Key Methods for Enhancing Memory
Endnotes
My book, ULTRALEARNING, is now available.
How does your memory work? How can you remember more? Prevent
forgetting?
ese questions lie at the heart of anything you’ll ever learn, do or
experience. Your memory impacts everything you do, from how well you’ll
do in school, your career, life and even your sense of self and happiness.
is is a topic that has long fascinated me, and I’ve written a lot about
memory previously on this blog. However, I wanted to create a guide that
would combine and integrate everything we know scienti cally about
memory, and distill that knowledge into practical advice.
To do that I collaborated with Jakub Jílek, who has his masters in cognitive
science at University College London, and is currently studying for his PhD.
In this complete guide, we will cover everything you need to know about
memory, how it works and how you can improve it.
Why Memory Matters
What is memory? e general consensus is that memory is a multitude of

cognitive systems which allow us to store information for certain periods

of time so that we can learn from our past experiences and predict the

future.

Memory impacts every facet of our lives. e rst step to remembering

things better is to understand how your memory works.

ere are two basic kinds of memory – retrospective and prospective.

Whereas retrospective memory is about remembering what happened in
the past, prospective memory is about reminding yourself to do something
in the future. Without prospective memory, you would not remember to go
to work in the morning and you would forget to set your alarm clock in the
evening.
One way to divide up retrospective memory is in the kinds of things it
stores. A big di ference here is between implicit and declarative memory.

Implicit memory is essentially skill memory – the ability to do a task. If

your implicit memory failed, you would not be able to brush your teeth,
take a shower, drive your car or ride a bike. is kind of memory shows up
in our abilities, but we can’t always articulate what it is we know in words
and concepts.
Declarative memory, in contrast, is either memory for facts and meaning
(semantic) or memory for events (episodic). Without semantic memory,
you would not understand the content of what your colleagues or friends
were saying. Without episodic memory, you would struggle to recount your
day later to someone else.
Another way to examine memory is based on its duration. Working
memory (WM) manipulates and stores information for short periods of
time. Talking with your colleagues, discussing a point at a meeting and
planning your weekend would be entirely impossible without WM. In
contrast, long-term memory (LTM) serves as a long-term storage of
information. Almost all of our everyday activities depend on LTM, such as
remembering our way home or how to drive a car.
Here’s a quick recap of the di ferent types of memory:
Retrospective Memory — Remembering the past
…by Content
Declarative / Explicit
Semantic — Memory of facts and meaning
Episodic — Experiences you’ve had
…by Duration
Working Memory – What you can “keep in mind” at
the same time, to think, reason and solve problems.
Long-term Memory – Anything you remember
longer than a moment–what you ate for breakfast,
exam questions or your friend’s birthday.
Prospective Memory — Reminding yourself to do something in the
future
In this guide, we’re going to focus mostly on declarative, semantic
memories. is is covers most of the things you’re trying to “remember”,
like facts, dates, names and ideas.
 The Three Parts
Underlying All Your Memories
What does a ten-dollar bill look like (or a common banknote, if you’re not
American)? Do you think you’d be able to draw one? Although we see coins
and banknotes on a regular basis and therefore have virtually limitless
opportunities to learn their shape, few people could sketch one accurately.
Looking at something repetitively does not guarantee that we will
remember it later. Why can’t we draw a ten-dollar bill, yet we could
recognize it instantly if we saw one?

To unravel this mystery, we need to break the act of remembering things

into its atomic parts. ose parts are:
1. Encoding—the process of putting the information into your brain.
2. Storage—the process of keeping the information in your brain.
3. Retrieval—the process of getting the information out of your brain
when you need it.

Understanding these three functions is essential if you want to have a

better memory. Any attempt to improve your memory must either encode
the information better (or in a format you’re more likely to retrieve), store
the information better and longer, or retrieve it in the situation you need.
Let’s look at all three and see how we might be able to improve our
memories.
Encoding: Putting memories into the
brain
Encoding is a process of imprinting information into the brain. Without
proper encoding, there is nothing to store and attempting to retrieve the
memory later will fail. One way to improve encoding is simply to repeat the
information more times. Scientists who study memory call these
repetitions “rehearsals” of the information.
However, as the case with the bank note illustrates, many rehearsals do not
always mean you’ll successfully recall the information. What makes the
di ference between this case and the more familiar case of remembering
your phone number because you’ve had to use it a lot?
When you were trying to memorize the phone number, you did not merely
look at it repetitively. Instead, you deliberately tried to memorize it. You
may have read it to yourself several times. Maybe you attempted to recall it
from your memory, checked whether you were right and corrected yourself
accordingly. Perhaps you noticed that there were some relationships
between di ferent numbers (e.g. some numbers were the multiples or sums
of preceding or following numbers). In summary, you employed certain
cognitive strategies and processes. You still needed repetition, but
repetition was e fective only when used together with these strategies.
Similarly, if you want to have an e fective memory, the cognitive strategies
you use will make a big di ference on your ability to remember things later.
What kind of encoding strategies are most e fective?
Intention to learn: Does it matter how
much you want to remember?
In an intriguing study, scientists asked students to study lists of words.[1]
One group was explicitly told to memorize the words (with a warning that
there would be a test later), whereas the other two groups were asked to
either sort the words into categories or to simply arrange them into
columns (unaware that there would be a test later).
One would expect that the students who deliberately studied the words
would perform better than the categorizing and arranging groups, who did
not make such e fort. However, this was not the case. e categorizing and
memorizing groups performed equally well in the test, whereas the
arranging group performed worse than the other two groups.
Why didn’t the students intention to learn make much di ference?
In brief, the researchers found that the students who were told to
memorize categorized the words in the same way as the students who were
explicitly told to do so. By categorizing, the two groups e fectively
processed the words more deeply than the students who simply arranged
them. As a result, they encoded the words more strongly than the
arranging group and achieved better test results.
is experiment shows that the mere intention to learn something is not
what makes you remember it later. What matters is what you do with the
material, i.e. what speci c strategies you use to process it, rather than how
much you want to memorize it.
Let us take a look at these strategies.
Depth of Processing: Why how you process information determines
how much you’ll remember later
Scientists believe that one of the critical factors determining the success of
encoding is the depth of processing. e deeper you process the to-be-
learnt information, the more likely it is to stick. What exactly is ‘deep’
processing?

In essence, deep processing focuses on the meaning of the information.

Try this demonstration:[2] Read the following list of words 3 times. A ter you
have read the list, try to recall as many words as possible, minimizing your
window or otherwise covering up the words on the screen:
cabbage, table, river, shirt, gun, square, iron, dentist, sparrow, mountain,
hand, granite
How many words did you remember? Now try the same thing with another
list:
pink, green, blue, purple, apple, cherry, lemon, plum, lion, zebra, cow,
rabbit
How many words did you remember? Try it once more with the following
list:
thread, pin, eye, sewing, sharp, point, prick, thimble, haystack, thorn,
hurt, injection
How many words did you remember now? It is quite likely that you
remembered most items from the second and third lists and the fewest
items from the rst list. As you may have noticed, the second list consists
of items grouped into categories (color, fruit and animal), whereas the
third list contains items which are related to the word ‘needle’. In contrast,
the rst list consists of completely unrelated items.
e reason why it was easier to memorize items from the latter two lists is
that the items were meaningfully connected – they were subjectively
(consciously or unconsciously) organized into a speci c category or related
to a common concept. Giving meaning to information is bene cial as it
harnesses the process of spreading activation (which we’ll cover in the
section on “Spreading activation”).
e main implication of this study is that structured information is much
easier to encode to memory than disorganized information.[3]
erefore, it is extremely useful to impose a good structure on your notes.
e structure can take many di ferent shapes – hierarchical, low-based,
mind-mapping or anything else that you nd useful. What matters is that
the particular technique helps you organize the study material in an easy,
clear and understandable way. (we will cover these note-taking techniques
in detail in a separate article on learning)
Categorization and structuring are not the only ways that you can give
meaning to information. A powerful technique that substantially improves
memorization is self-explanation.[4] Self-explanation simply consists of
asking yourself questions about the study material as you study:
How does this concept related to the other?
What are the implications of this concept for…?
Why does it make sense that…?
What are the steps that I must take to solve this problem?
A very e fective way to make yourself process information deeply is to
explain the study material in your own words. If you are to explain a
concept using di ferent words than those used by the textbook or the
lecturer, you rstly have to process and understand its meaning and logical
connections with other concepts, which e fectively boosts encoding by
stimulating deep processing. is stands in stark contrast with the
situation where you simply re-read the textbook or lecture notes, which
constitutes only super cial processing and does not lead to e fective
encoding.

When taking notes, make sure that you do not copy the words of your
textbook and lecturer verbatim. Instead, try to use your own words as
much as possible. Researchers have shown that typing notes on a computer
encourages copying information verbatim (even if students are explicitly
instructed to use their own words), unlike writing notes by hand. As a
consequence, students who take notes on a computer underperform in
tests compared to students using handwriting.[5]
Transfer-appropriate processing: The trick to acing your exams
Imagine yourself learning how to ride a bike. You could buy a 200-page
long book on cycling and memorize everything perfectly. If you were to sit
a written test, you would ace it.
Now imagine that you were to actually ride your bike. What do you think
would happen? e chances are that you would crash as soon as you got on
your bike.

Although you knew everything you could about cycling, a key element was
missing. e reason for the crash is that the cognitive processes used
during during encoding did not match (=transfer appropriately to) the
processes needed during retrieval. To remember e fectively, the processes
used during practice need to correspond with the processes during use.
As an illustration, consider the following study:[6]
Researchers asked students to either read aloud a list of words (super cial
processing), or to generate these words from their antonyms (deep
processing). e students were later asked which words they could
remember (free-recall) or to ll in missing letters in words (fragment
completion).
One would expect that the super cially-processing students would
underperform in both tests (because deep processing is generally better
than super cial processing – see previous section). However, this was the
case only for the free-recall test. Surprisingly, in fragment completion, the
super cial-processing group was better than the deep-processing group.

What could explain this surprising result?

e reading group processed the words perceptually, while the generating
group processed them semantically – they had to retrieve from memory
words with a particular meaning. Whereas perceptual processing matched
the processing needed by the perceptual task, semantic processing
matched the processing needed during recall (note that words are stored in
semantic memory based on their meaning).
e main implication of this study is that although deep processing is
extremely bene cial for memory, it may not be enough to fully optimize
your test performance. To further improve your results, it is important that
you practice with similar processing that will be required during the test.
ink hard about how you will be tested on the information you need to
remember. Will it be multiple choice tests? Essay questions? Applied in real
life problems? en make your practice match the situation where you use
it. Mismatched practice is a major cause of poor memories–they simply
aren’t encoded in a way that is useful.
If your exam will consist of writing an essay, an excellent strategy is to do
your reading with pre-reading questions.[7] Pre-reading questions force you
look for arguments and evidence in order to answer the questions, which
are precisely the processes that you will need during your essay-type exam.
However, it turns out that transfer-appropriate processing is only one
consideration that matters for memory. is is because some encoding
strategies are generally better than others, regardless of whether they
match the test format or not. In fact, one speci c encoding strategy
dominates almost all other strategies. is strategy is called “recall” and is
discussed later.

State-dependence: How your physical/mental states drive your

memory
Imagine the following scenario: You have to prepare for tomorrow’s test
but your friend has a birthday party tonight. You decide to go to the party
and end up having a few alcoholic drinks. When you come back home, you
are quite drunk but you study for the test anyway. e next morning you go
to school to sit the test. Would you be better o f taking a shot or two of an
alcoholic beverage before the test or is it a better idea to refrain from
drinking any more alcohol?
Setting aside the fact that you would likely not be admitted to school in a
drunk state, science has an astonishing answer: In order to improve your
performance, you should top up alcohol to approximately the same level
you had during revision (this was actually shown in a study[8]).
What could explain this surprising result?
Research has show that our memories are state-dependent. e more
similar are our mental, physical and chemical states between encoding and
retrieval, the more likely we are to successfully remember. Memory relies
on a chemical process through which new connections (pathways) are
formed and strengthened between neurons. If you study in a particular
state, the memory trace is encoded with brain activity in luenced by this
state and becomes to some degree dependent on its reinstatement.
State-dependence of memory has been found for all kinds of drugs and
medications, including nicotine[9], ca feine[10], cannabis[11], Ritalin[12] or anti-
histamines[13]. If you are on medications such as Ritalin, it is therefore a
good idea to keep the same dose during both revision and testing. If you
cannot drink co fee or smoke cigarettes while you are taking a test, you had
better avoid these drugs during revision as well. Moreover, it is important
to realize that, the majority of drugs have well-known detrimental e fects
on memory, especially alcohol[14] and cannabis[12]. You stand the best
chance of passing your test if you both revise and take the test while you are
sober.
State-dependence of memory applies to other states as well. If you study in
a good mood, you are likely to perform better in a test if you are also in a
good mood (and the same applies to other moods)[15]. Similarly, if you study
while standing up/doing aerobic exercise, you are more likely to remember
the material if you are also tested while standing up/doing aerobic
exercise[16].
One approach to overcome state-dependence of memory is to try to study
in the same state that you will be in during the exam. You could, for
instance, revise sitting at a desk, while simulating stressful feelings (e.g. by
timing your answers), assuming that these conditions will be the same
during the exam.
However, this is not always possible. An alternative approach is to study in
various mental and physical states. e logic behind this is that you never
know what kind state you will be in during your exam therefore it is best to
make your memory independent of any particular states. For example, you
could revise both when you have a lot of energy and when you are low on
energy. Also, it is a good idea to study regardless of the mood you are
currently in.

Context-dependence: Why your environment matters

Consider an everyday situation: You get up from your desk to have a cup of
tea. Once you arrive in the kitchen, you forget what you wanted. However,
when you get back to your desk, you suddenly remember.
Scientists have discovered that memories are heavily context-dependent.
Context is essentially anything that is present during encoding (for
instance the environment we are in). Our brains seem to encode the
context as a part of the memory trace as if taking a snapshot of everything
that is around us at the moment of creating the memory.
Successful retrieval of the memory trace then depends to some degree on
the re-activation of the context in which it was encoded. Since the
intention to have a cup of tea was encoded with the context of the kitchen,
coming back to the kitchen re-activated the intention to have a cup of tea.
To combat context-dependence, you can adopt the same two approaches
used for overcoming state-dependence. e rst approach would be to
emulate the environmental context of the test. For instance, you could
revise in a quiet/noisy environment depending on where your exam will be
situated. You may also consider revising together with a friend or two to
get used to being distracted by other people in the examination room. An
even better idea would be to revise in the classroom where you will be
taking the test.
e second approach would be to revise in as many di ferent contexts as
possible. Studies have shown that students who revise in many di ferent
rooms prior to their test perform better than those who study in one room
only (with a 30% improvement in test performance)[17].
Since the environmental context keeps changing, the information
e fectively becomes context-independent. In other words, you teach
yourself how to retrieve the studied material in any kind of circumstance,
which is extremely useful given that fact that you o ten cannot predict the
exact circumstances you will face during the exam.
e context of study need not be only environmental. e particular
questions and practice tests you use also become the context that is
encoded with your study material. erefore, the more questions you
practice on for a given concept, the more neural connections the brain has
to generate between di ferent contexts and the target concept. e more
routes the brain has built, the easier it is to retrieve the concept later. is
is because retrieval becomes less dependent on the particular starting point
– the type of question asked or its particular wording.
e impact of this kind of context-sensitivity is particularly important
when creating lashcards. If the question side of your lashcard contains
irrelevant information, or information that won’t be present when you
really need to remember, you may not be able to recall it when you need it.
Consider the following lashcards:
Q: How do you say “again” in Chinese, but only for actions that you will
repeat in the future? Like asking someone to do repeat something they just
said…
A: 再
Compare that to:
Q: Again (future)
A: 再
e former has so much more context, that you may memorize the pairing
only with this context (which may be missing when you need to think about
the term).
For this reason, it is better to either ask yourself questions with as little
context as possible (and thus maximum di culty) or to ask many di ferent
questions, with di ferent contexts, so the same context isn’t required for
successful retrieval.
 Storage: Keeping
memories in the brain
Once you’ve encoded information, you now need to store it. Unfortunately,
forgetting is a major part of how our brains work. Most of us can’t
remember what we had for dinner Tuesday, three weeks ago. However, we
can all remember our rst kiss.
Forgetting can be caused by two di ferent processes. e rst is a failure of
storage—the information might be forgotten because our brain loses it
over time. e second is a failure of retrieval—the information might be
“in” there, but we can’t access it. Experimentally, it’s very di cult to tell
these two apart, but since they are separate processes, we’ll consider each
separately as we look at how memory works.

e progression of forgetting was originally studied by the famous

experimental psychologist Hermann Ebbinghaus. Ebbinghaus discovered
that forgetting follows an exponential decay function. Whereas in the rst
hours a ter study there is a rapid drop in the number of items successfully
remembered, there is relatively little forgetting from the 2nd day onwards.
Although the exact shape of the forgetting curve is likely individual and
depends on many factors related to the study material (e.g. how
easy/di cult/interesting the material is or how well it was encoded), it
approximates an exponential curve, with rapid forgetting in the beginning
and less forgetting in later periods of time.
Why do we forget information in the rst place? What can we do
counteract the process of forgetting?
Trace decay: Do your memories fade with time?

e most basic explanation of forgetting has to do with the passage of

time. Our memory traces are stored in living tissue, which inevitably
changes over time. It is a well-known fact that connections between
neurons deteriorate over time and as a consequence, the memory traces
stored within these connections decay.
In addition, there is another possible cause for the decay of our memories.
Surprisingly, making new memories may be part of the reason we forget.
As new memories are formed, new neurons are produced in the
hippocampus (the memory hub), which changes its structure and patterns
of connections.[18] As a consequence, older memories are more di cult to
retrieve.
erefore, if you need to retain old information in memory (for a nal
exam for instance), it would be a good idea to revise it again while you are
studying something new, otherwise the old information might be
superseded by the new information.
We recommend that you establish a schedule of revision of the old material
(even a couple of minutes every day should su ce) that is interspersed with
the study of the new material. is is especially important if you study
several di ferent subjects/topics within one day because you memory
cannot yet bene t from protective sleep-induced consolidation processes.
ere is another reason why studying new things can impair the retrieval of
old things and we discuss it in the section “interference”.

Consolidation and sleep: Let biology do the work for you

Learning does not nish with the end of studying. For a memory trace to
become permanently established in our long-term storage systems,
structural biological changes must take place in brain tissue. New
connections between neurons must be formed and rmly established.
ese changes are not immediate and take time. In scienti c terms, the
mechanism through which recent memories become permanent memories
is called ‘consolidation’. Although some consolidation occurs during
wakefulness, the primary time for consolidation is sleep.
Have you ever studied for an exam through the night? If so, did you
perform as well as in other exams for which you enjoyed a full night of
undisturbed sleep? Probably not. Unsurprisingly, researchers have found
that sleep deprivation impairs memory consolidation and undermines
learning.[19] In fact, sleep deprivation (before or a ter learning) can worsen
performance in a declarative memory test by as much as 20-50%[20][21].
Moreover, prolonged sleep deprivation has permanent damaging e fects on
memory.[22]
On the other hand, you can use sleep as a powerful aid in-between your
study sessions. You may have heard of ‘power naps’ – short periods of sleep
used to refresh energy. ere is now robust evidence to recommend naps.
[23]
Napping during the day will protect your memory from trace decay (see
section ‘Trace decay’) and interference (see section “Interference”) due to
sleep-induced consolidation processes. In other words, you will forget less
of the subject studied than if you stayed awake for the same amount of
time.
If you decide to give napping a try, it is important to be aware of the
di ferent stages of sleep. Napping for the maximum of 20 minutes is
e fective for restoring your energy, however, it is not enough to reach
deeper stages of sleep during which consolidation occurs.
In order to boost your memory, you need to sleep for at least 60 minutes.
However, napping for 60 minutes has the downside of leaving you in a
groggy state for about 30 minutes a terward (because you wake up in the
middle of deep sleep). erefore, it is best to sleep for a full 90 minute cycle.
A ter that, you will both feel refreshed and your memory will bene t from
consolidation. Another good option that has been found to be e fective is to
schedule your study session to the evening right before sleep.

Interference: Does learning new things block your old memories?

Do you remember what you had for dinner two weeks ago? Now choose
your favorite trip from a couple of years ago. How much do you remember
from that trip? e chances are that you do not remember what you had for
dinner but you do remember something about your trip, although it took
place much earlier than the meal.
is example shows that forgetting is not simply memories decaying with
time. Our memories crucially depend on cues. A cue is essentially anything
(such as a physical object, situation, time period, word, question, concept,
etc.) which is paired with a memory trace and which must be activated for
the memory trace to be retrieved.
If we pair the same cues with multiple memory traces then it will be
di cult to retrieve one particular trace because once the cue is activated,
the activation will spread to all paired memory traces at once and these will
compete for entry to consciousness.
Coming back to the example above, if you usually dine in the same place,
many di ferent meals will become associated with the same cues (the
dining environment). erefore, it will be hard to retrieve the speci c meal
that you enjoyed a week ago. In contrast, you probably have not been on the
same trip many times before, therefore it is easier to remember its details
because they context of the trip is not paired with any other memories.
e disruption of memories by other memories which are paired to the
same cues is called “interference”. You may have experienced interference
yourself if you ever studied a second language. Interference may have
caused you to be unable to retrieve vocabulary from one language. Instead,
vocabulary from the other language popped to your mind. In this case,
interference did not necessarily cause a loss of memory, but the memory
trace became blocked thus temporarily inaccessible.
Research has found that the only way to overcome blocking interference is
by making conscious e fort to recover the correct memory trace (and have
patience as this may take some time). Interference may, however, also
cause a permanent loss of memory. Scientists who study memory call this
the retrieval-induced forgetting e fect (RIF).
As a demonstration, consider the following experiment: Students studied
10 geographical facts about each of 2 islands (A and B).[24] ey
subsequently practiced retrieving 5 out of 10 facts for island A. A terwards,
their knowledge of these facts was tested.
What do you think happened to students’ memory about island A?
Unsurprisingly, retrieval practice boosted retention for the 5 facts that were
practiced (the percentage of correct answers was greater than for island B).
However, it also worsened the memory for the 5 facts about island A that
were not practiced (again compared to island B). What caused this e fect?

e island A serves as the context cue for information about island A,

whereas island B serves as the context cue for information about island B.
When the 5 facts about island A were retrieved from memory, their
connection with the context cue was strengthened and the connection of
the remaining 5 facts with the context cue was weakened (see “Testing
e fect”).
e main implication of this study for learning is that selective practice
testing substantially boosts performance for the practiced items but can
also worsen the performance for the unpracticed items. How can we
combat forgetting caused by interference?
One way we can overcome interference is by making it explicit. If there are
concepts that you get mixed up frequently then put them side by side and
re-study them at the same time.
e general idea is that whatever you are studying, it is good practice to
make di ferent concepts as distinctive as possible.[25] is forces your brain
to encode them as dissimilar memory traces. You can achieve this by
stressing the di ferences between di ferent concepts from your study
material (by comparing and contrasting, for instance).

Another e fective strategy is to integrate the concepts.[26] For instance, if

you are memorizing the members of a particular animal/plant family, then
try to nd all possible relations between the members. When you’re later
retrieving these members, they will no longer compete for access to
consciousness as they will be encoded closely together in an integrative
manner. Instead of one concept blocking the other, they will be retrieved
simultaneously.
Scientists have found that our study goals also impact on how well we
overcome interference.[27] Students who focus on comparative performance
(how well they do compared to other students) tend to use super cial
processing (do not look for relations among concepts), whereas students
who aim for mastery tend to use more deeper processing, such as
establishing connections between di ferent concepts.

Spacing e fect: Study less, remember more

 e spacing e fect is undoubtedly one the most important discoveries in
the science of memory. e general idea of spacing is that to achieve the
same performance at a given test, you need substantially less time overall
to memorize something if you spread your study into multiple sessions as
opposed to if you study everything in a single session. As a demonstration,
consider an experiment that the famous psychologist Hermann
Ebbinghaus performed on himself:[28]
Ebbinghaus studied 2 lists of words on two following days. On the rst day,
he spent 1 minute studying list A and 7.5 minutes studying list B. On the
second day, he had to spend another 20 minutes on list A to memorize it
perfectly. However, he needed only 7.5 extra minutes to memorize list B.
By spacing the sessions more equally for list A, Ebbinghaus managed to
save himself about ¼ of total time spent on studying. In a similar way, if it
takes you 3 hours in a row to prepare for a test, you may need less than 2
hours in total if you divide the time into 2 or 3 equally-sized sessions
spread across several days.
In general, the more you spread your sessions, the less overall time you will
need. Scientists recommend that the spacing distance should be about 10-
20% of the test delay.[29] If your test is in 10 days, you will bene t most from
studying once a day. If your test is in 6 months, you should study every 20
days.
Spacing not only substantially saves time, it also boosts long-term
retention. Each study session that is followed by immediate or delayed
sleep provides another opportunity to consolidate the studied material (we
cover consolidation processes above). Furthermore, spacing can give you
more opportunities to associate the study material with more states and
contexts (physical, mental, environmental), which makes it easier to
retrieve it in the future (see state-dependence and context-dependence).
Most importantly, spacing e fectively reveals to your brain that forgetting is
taking place – a crucial signal that is not available in massed practice (= if
you study for many hours within a single session). e signal of forgetting
has been found to automatically encourage more e fective encoding
strategies in students.[30]

e spacing e fect is a robust phenomenon in psychology and is extremely

e fective for a wide range of study materials (including foreign language
vocabulary, math problems, spatial information).[31]
e key idea behind spaced repetition is that revision of a particular
concept should be done right before the concept is about to be forgotten in
order to achieve maximum time-e ciency and length of retention. As
there is no formula that could calculate this exactly, you need to experiment
with yourself. For practical reasons, it is probably best to use a specialized
so tware that uses a spacing algorithm.
Although there is currently no comparison of the e fectiveness of di ferent
spaced-repetition algorithms, popular choices include Anki, Quizlet or
Supermemo. We recommend that you experiment with a program that is
free on most platforms, explicitly states the algorithm used and allows you
to tweak it (our favorite pick is Anki).
Bear in mind that the spacing e fect does not continue inde nitely. For
example, ten 1-minute sessions will likely not be better than one 10-minute
session. is is because starting each learning sessions carries some costs
with it (you have to start to focus, ‘load’ the material into working memory,
etc.). For this reason, complex subjects may require longer study sessions
for learning to be e fective. For instance, it may be better to have three 1-
hour sessions per week for your engineering class than six 30-minute
sessions. It all depends on the subject studied.
Retrieval: Accessing Memories in Your
Brain
Retrieval is the mechanism of accessing information stored in memory.
Successful retrieval of a memory trace hinges on its associations with cues.
A cue is anything that is connected to the memory trace (physical object,
situation, time period, word, question). Scientists believe that memories
are retrieved through the process of ‘spreading activation’. Once a cue is
activated in the brain, the activation spreads from the cue to the target
memory. A single memory trace can be connected to an in nite number of
cues. If none of the relevant cues is activated, the memory trace cannot be
retrieved, even though it may be well stored in memory.
As an example, try to remember the capitals of the following countries
while covering the rest of the page:
South Korea
Syria
Denmark
Colombia
Afghanistan
ailand
Venezuela
Turkey
Could you remember all of the capitals? Do you feel that you know their
names and may be able to remember them later? You may be experiencing
the tip-of-the-tongue phenomenon: you know that you know something
but still can’t remember it. Now try the same exercise with a little help:
South Korea S
Syria D
Denmark C
Colombia B
Afghanistan K
 ailand B
Venezuela C
Turkey A
Did you remember all of them now? You most likely remembered more of
them. is is because the starting letter functions as a suitable cue that is
connected to the capital’s name. When the cue is provided, the memory
trace storing the capital’s name becomes automatically activated.
How does the process of retrieval function and what are its implications for
learning? What can we do to maximize our chances of successfully
retrieving information?
Spreading activation: Gain quick access to your memory

Imagine that you are preparing for a test on all capital cities in the world
(given a country, you have to state the capital). If you used the most basic
learning strategy, you would simply learn to match each country with its
corresponding capital. Alternatively, you could look at pictures taken from
these capitals, maybe watch short videos of people speaking the country’s
language, visiting the capital sights and so on.
Which strategy would be more e fective?
e rst strategy would be called ‘shallow’ processing because you would
not be giving the information any additional meaning. You would be
making only one connection between the cues (countries) and the target
memories (cities). e second strategy would be called ‘deep’ processing as
you would be drawing many connections – between the capital, the
country, its people and its sights, and so on.
If you adopted the second strategy, you would be making use of the
structure of your memory. Memory is believed to operate on the principle
of spreading activation.[32] If you encounter one concept (a country), the
neural pathways representing this concept are activated in your brain. As a
consequence, nearby neural pathways encoding closely related concepts are
also activated. Images and ideas come to your mind. For instance, when
someone says France, you may visualize the French lag, the Ei fel tower,
French wine, cheese, etc. is process continues to further and further
concepts until you start thinking about something else.

If you previously connected these images with both France and Paris, the
cue France will activate additional related cues (Ei fel tower, wine, cheese,
etc.) and together they will generate more activation than France alone,
which spreads to the connected memory trace – Paris. Conversely, if you
studied the France-Paris pairing super cially, seeing the word France
might not generate su cient activation on its own to trigger the memory
trace of Paris.
e main implication of spreading-activation is that in order to maximize
the chances of remembering new concepts, you should try to make as
many connections as possible between the new concepts and what you
already know. If the target concept is connected to many other concepts,
the chances of it getting activated (and retrieved) are generally much
higher than if it only has few connections.

Retrieval failure: What to do if you get stuck in an exam

Retrieval failure or the failure to remember a memory trace can have
multiple reasons. One reason can be a lack of attention during study.
Researchers have found that students who do a secondary task while
studying underperform in a later test by as much as 30-50% compared to
students who focus on one thing at a time.[33] ese results suggest that
multi-tasking – doing of many activities at once – is particularly harmful
to learning.
Another reason for retrieval failure is an insu cient number of activated
cues. Cues are pieces of information which are connected to the target
memory trace and which must be activated for the memory trace to be
retrieved (for more more detailed explanation of context cues, see sections
on Retrieval and Interference).
Activation from the cue to the target memory trace spreads faster if more
cues are activated simultaneously. As a demonstration, try the following
little experiment:
ink of the name of any kind of supernatural creature.
Now think of a supernatural creature that rhymes with ‘post’.
Did you think of ‘ghost’ on the rst time or only a ter the second question?
e rst question gave you only one cue, which led to multiple possible
target memories. For instance, you may have thought of fairies, gnomes,
angels and the like. e second question gave you 2 cues. ese two cues
jointly generated enough activation of the word ‘ghost’ that far surpassed
the activation of other concepts. erefore, the word ‘ghost’ was retrieved.
Consider another example from everyday life: You decide to return a book
to the library while sitting at the kitchen table. Later, as you pass the library
on your way from school, you forget to return the book. However, when you
come back home and see the kitchen table, you suddenly remember.
Memory traces are encoded together with the context present at encoding.
Although the library would surely be a far more relevant context cue than a
kitchen table, it was the kitchen table – not the library – that was present
during encoding and thus encoded with the intention.
When you are forming an intention, it is very useful to imagine yourself
doing the desired action in the desired context as vividly as possible. For
instance, visualize yourself passing the library, entering the building and
returning the book. Focus on the details – which objects/buildings are you
likely to notice on your way round the library? By doing this, you connect
the cue of passing the library (and the surrounding objects) to the
intention. When you later pass the library, it will automatically trigger the
intention.
If you are struggling to remember an important concept during your exam,
you need to activate as many connected cues as possible. For instance, try
visualizing yourself in the context of studying. Be as vivid as possible –
imagine yourself with an open textbook, taking notes, sitting at your desk.
Simply imagining the context of encoding can be helpful
to generate enough activation to successfully retrieve the memory trace.
Also, try to remember the details of the context in which you studied the
particular concept you are struggling with (such as what page it was in the
book, what other concepts you studied before and a ter this concept). Note
that for successful retrieval it is important to activate the connected cues.
Since our memories work like snapshots (everything that is present during
encoding is encoded together with the memory trace), these cues can
either be relevant (such as related concepts) or even completely irrelevant
(such as the time of day or even what you had for lunch around the time of
study, etc.).
Practice tests: The most powerful technique for boosting memory

Scientists have found that regardless of type of test/exam you are going to
take, you stand the best chance of succeeding if you revise with practice
tests.[34] As a demonstration, consider the following experiment:[35]
Students attended a 20-minute Statistics lecture, which was divided into 4
equally long sessions. A ter each session, the rst group took a practice test
(without feedback), the second group had to re-study the lecture material
and the third group performed mental arithmetic. All students were
assessed with a nal test a ter the lecture.
Although the re-reading group had more exposure to the material, their
nal test performance was substantially worse than that of the testing
group (by 30%). Moreover, students who re-studied the lecture material did
not perform any better than students who did a completely unrelated
arithmetic task.
A wealth of research has shown that testing is more e fective at improving
retention and test performance than re-studying, even if no feedback is
provided. is is called the ‘testing e fect’. However, not all tests are created
equal. You will greatly bene t from practice testing only if you revise with
tests using a particular retrieval mode – recall.
Recall is a way of retrieving a memory trace when you do not see the correct
answer and do not have any options to choose from. Example recall
questions could be: “What is the population of Canada?”, or “What is the
German word for Monday?”
e opposite of recall is recognition, which is a way of retrieving a memory
trace when you see the correct answer or a set of options that includes the
correct answer. Example recognition questions could be: “Is the capital of
Canada Ottawa or Montréal?”, or “Is the German word for Monday
Dienstag or Montag?”.
Regardless of how your knowledge is tested in the end, recall testing is
vastly superior to all other learning methods based on recognition. One of
the reasons for this is that it automatically encourages deeper processing of
the study material. In the study above, the testing group took increasingly
elaborate notes as the sessions progressed compared to the other groups
(without being consciously aware of it).
e most important reason is that recognition is easy for the brain because
it knows exactly which concept it must retrieve from memory. If you re-
read your notes, you are in fact asking your brain: do I know this? Does it
sound familiar? If you have already studied the notes at least once before,
you are in e fect telling your brain: “I’ve seen this before, there is no need to
make any further e fort”.

In contrast, recall is e fortful because the brain has to gure out which
target memory is to be retrieved. If you test yourself with recall (questions
with no hints or answers to choose from), your brain has to reconstruct the
pathway from the question to the target concept. In this manner, the
pathway is strengthened (or new pathways are built) and as a result, the
concept becomes more easily retrievable later.
Note that the recall has to be successful – unsuccessful recall does not
strengthen the memory trace. e best time to revise concepts is therefore
just before they are about to be forgotten (see our section on spacing
e fect).
Virtually all methods commonly used in studying engage primarily
recognition processes, such as reviewing (re-reading), highlighting or
open-book summarizing. It comes as no surprise then that these learning
methods have been shown to have little to no utility in improving retrieval
success and test performance. Other methods using deeper processing
such as self-explanation are far more useful than re-studying but still not
as e fective as practice testing.[36]
However, it is not the case that re-reading has no value whatsoever. Re-
reading is useful inasmuch as it is used together with practice testing. It is
de nitely a good idea to selectively re-study the concepts which you cannot
recall.
Also, it is important to re-study material during practice testing as a form
of feedback. Although practice testing without feedback is very e fective on
its own at improving memory, if errors go uncorrected, they build up over
time and become more and more rmly lodged in memory. For this reason,
feedback is an essential complement to practice testing that substantially
enhances its e fectiveness. It does not matter whether feedback is
immediate (straight a ter each question/problem) or delayed (a ter the
study session).[37]
Finally, do not forget that practical usage your knowledge (such as doing a
real-world project) also amounts to a form of practice testing (and spaced
repetition) where you have to regularly retrieve your knowledge and skills
from memory. A fair amount of real-world practice can be superior to
extensive theoretical study.
 Summary of Key
Methods for Enhancing Memory
Let’s recap what we’ve learned!
Memory has three parts: encoding, storage and retrieval. All three need to
function successfully to remember what you need to.
To encode information better:
Process information deeply. Pay attention to the deeper meaning,
make connections to what you already know and paraphrase rather
than take notes verbatim.
Intention doesn’t matter so much. Trying to remember more
doesn’t make a di ference if you use the same cognitive strategies!
Match your practice and studying time to how you’ll eventually use
the information. Greater overlap means you’ll remember more later.
If possible, align your state and context when you’re encoding the
information to when you need to remember it. If this isn’t possible,
study in more environments/situations to make your memories
more robust!
To store information better, you need to be aware of how you forget. Here
are the main possible causes of forgetting:
Trace decay. is occurs when memories get old, or new knowledge
overwrites the old data. Refresh important information on a
schedule so it doesn’t get lost!
Interference. is happens when a new memory blocks an old one
(e.g. you can’t remember the Spanish word for water anymore
because you learned the French one.) Alternatively, it can happen
when old memories make learning something new harder.
How can you store your memories better?
Get enough sleep! Short naps can recover energy. Longer naps (60+
minutes) can enter the phase of sleep where memory consolidation
happens. Ending your naps on a full sleep cycle can prevent
grogginess. Of course, getting enough sleep at night is essential!
Space out your practice. Done properly, you can get the same
 memory strength with 20-30% less time by spacing properly.
Finally, you need to retrieve the memories in the situations that need them.
How can you do this?
More connections help. Memories are likely accessed by spreading
activation, so if you think of related items, that can help you retrieve
something di cult.
Plan ahead and visualize the context you’ll need to retrieve
something in when you study.
Practice testing is the single most-e fective technique you can use!
Practice recall, not just recognition. Harder recall creates stronger
memories.
 Endnotes
1 Mandler, G. (1967). Organization and memory. In K. W. Spence & J. T. Spence (Eds.), e
psychology of learning and motivation: Advances in research and theory. (Vol. 1, pp. 328–372).
New York: Academic Press.
2 Baddeley, A. D. (2015). Memory (2nd ed.).
3 Bower, G. H., Clark, M. C., Lesgold, A. M., & Winzenz, D. (1969). Hierarchical retrieval schemes in
recall of categorised word lists. Journal of Verbal Learning and Verbal Behavior, 8, 323–343.
4 Bisra, K., Liu, Q., & Nesbit, J. C. (2018). Inducing Self-Explanation : a Meta-Analysis. Educational
Psychology Review, (Siegler 2002).
5 Mueller, P. A., & Oppenheimer, D. M. (2014). e Pen Is Mightier an the Keyboard : Advantages
of Longhand Over Laptop Note Taking.
6 Jacoby, L. L. (1983). Remembering the data: Analyzing interactive processes in reading. Journal of
Verbal Learning & Verbal Behavior, 22(5), 485-508.
7 McCrudden, Matthew T. “Do Speci c Relevance Instructions Promote Appropriate Transfer
Processing?” Springer Science & Business Media 4 Nov. 2010: 865-79. Web.
8 Goodwin, D. W., Powell, B., Bremer, D., Hoine, H., & Stern, J. (1969). Alcohol and recall: State-
dependent e fects in man. Science, 163(3873), 1358–1360.
9 Swanson, J.M., Kinsbourne, M., (1976). Stimulant-related state-dependent learning in hyperactive
children, Science, 192(4246), 1354–1357.
10 Kelemen, W. L., & Creeley, C. E. (2003). State-dependent memory e fects using ca feine and
placebo do not extend to metamemory. e Journal of General Psychology, 130(1), 70–86.
11 H. Rickles, W., J. Cohen, M., A. Whitaker, C., & E. McIntyre, K. (1973). Marijuana induced state-
dependent verbal learning. Psychopharmacologia, 30, 349–354.
12 Swanson, J.M., Kinsbourne, M., (1976). Stimulant-related state-dependent learning in
hyperactive children, Science, 192(4246), 1354–1357.
13 Carter, S. J. and Cassaday, H. J. (1998), State‐dependent retrieval and chlorpheniramine. Hum.
Psychopharmacol. Clin. Exp., 13: 513-523.
14 Nelson, T., McSpadden, M., Fromme, K., & Marlatt, G. (1986). E fects of Alcohol Intoxication on
Metamemory and on Retrieval from Long-Term Memory. Journal of Experimental Psychology,
115(3), 247-254.
15 Eich, E., Macaulay, D., & Ryan, L. (1994). Mood dependent memory for events of the personal
past. Journal of Experimental Psychology: General, 123(2), 201–215.
16 Miles, C., & Hardman, E. (1998). State-dependent memory produced by aerobic exercise.
Ergonomics, 41(1), 20–28.
17 Smith, S.M. (1984). A Comparison of two techniques for reducing context-dependent forgetting.
18 Frankland, P. W., Köhler, S., & Josselyn, S. A. (2013). Hippocampal neurogenesis and forget- ting.
 Trends in Neurosciences, 36, 497–503.
19 Peigneux, P., Laureys, S., Delbeuck, X., & Maquet, P. (2016). Learning brain. e role of sleep for
memory systems. Neuroreport, (May).
20 Tilley, A. J. (1981). Retention over a period of REM or non-REM sleep. British Journal of
Psychology, 241–248.
21 Drummond, S. P. A., Brown, G. G., & Gillin, J. C. (2000). Altered brain response to verbal learning
following sleep deprivation. Letters to Nature, 304(1997), 655–657.
22 Alhola, P., & Polo-Kantola, P. (2007). Sleep deprivation: Impact on cognitive performance.
Neuropsychiatric disease and treatment, 3(5), 553-67.
23 Mcdevitt, E. A., Sattari, N., Duggan, K. A., Cellini, N., Whitehurst, L. N., Perera, C., … Mednick,
S. C. (2018). e impact of frequent napping and nap practice on sleep-dependent memory in
humans, (June), 1–12.
24 Macrae, C. N., & MacLeod, M. D. (1999). On recollections lost: When practice makes imperfect.
Journal of Personality and Social Psychology, 77(3), 463–473.
25 Reed Hunt, R. (2013). Precision in Memory rough Distinctive Processing. Current Directions in
Psychological Science, 22(1), 10–15.
26 Anderson, M. C., and McCulloch, K. C. (1999). Integration as a general boundary condition on
retrieval-induced forgetting. J. Exp. Psychol. 25, 608–629. doi: 10.1037/0278-7393.25.3.608
27 Ikeda, K., Castel, A. D., & Murayama, K. (2015). Mastery-approach goals eliminate retrieval-
induced forgetting: the role of achievement goals in memory inhibition. Personality & Social
Psychology Bulletin, 41(5), 687–695. https://doi.org/10.1177/0146167215575730
28 Ebbinghaus H. (1913). Memory: A Contribution to Experimental Psychology. New York: Columbia
University
29 Pashler, H., Rohrer, D., Cepeda, N. J., & Carpenter, S. K. (2007). Enhancing learning and
retarding forgetting: Choices and consequences. Psychonomic Bulletin and Review, 14, 187–193.
30 Karpicke, J. D., & Roediger III, H. L. (2008). e critical importance of retrieval for learning.
Science, 319, 966–968
31 Pashler, H., Rohrer, D., Cepeda, N. J., & Carpenter, S. K. (2007). Enhancing learning and
retarding forgetting: Choices and consequences. Psychonomic Bulletin and Review, 14, 187–193.
32 Schacter, D. L., Reiman, E., Curran, T., Yun, L. S., Bandy, D., McDermott, K. B., et al. (1996).
Neuroanatomical correlates of veridical and illusory recognition memory: Evidence from
positron emission tomography. Neuron, 17, 267–274.
33 Fernandes, M. A., & Moscovitch, M. (2000). Divided attention and memory: Evidence of
substantial interference e fects at retrieval and encoding. Journal of Experimental Psychology:
General, 129(2), 155–176.
34 Iii, H. L. R., & Butler, A. C. (2011). e critical role of retrieval practice in long-term retention.
Trends in Cognitive Sciences, 15(1), 20–27.
35 Szpunar, K. K., Khan, N. Y., & Schacter, D. L. (2013). Interpolated memory tests reduce mind
wandering and improve learning of online lectures, 110(16), 6313–6317.
36 Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving
Students’ Learning With E fective Learning Techniques: Promising Directions From Cognitive
and Educational Psychology. Psychological Science in the Public Interest, 14(1), 4–58.
37 Diego, S., Jolla, L., Diego, S., Jolla, L., Diego, S., & Jolla, L. (2007). Enhancing learning and
retarding forgetting : Choices and consequences. Psychonomic Bulletin & Review, 14(2), 187–193.
 My book, ULTRALEARNING, is now
available.

Being able to master hard skills quickly is essential to your work and life.
is book will show you how.
Combining stories of dramatic ultralearning feats with the detailed science
on how to learn anything e fectively, ULTRALEARNING will transform how
you work, live and learn.
Learn More:
https://www.scotthyoung.com/blog/ultralearning/