Psychology World — home
Blog

Sensory memory: the brief store before short-term memory

Sensory memory holds sights, sounds and touch for seconds. Types, durations, Sperling's test and examples.

Memory9 min read
By Psychology World

Sensory memory is the first and briefest memory store. It holds raw input from your senses for a fraction of a second to a few seconds, then passes a small part to short-term memory.

  • It is the gateway of the Atkinson-Shiffrin model. Unselected input fades and is lost.
  • It has one register per sense. Vision and hearing are the best studied.
  • Vision, called iconic memory, lasts about a quarter to half a second, with a usable trace up to about one second in Sperling's setup.
  • Hearing, called echoic memory, lasts about two to four seconds.
  • Touch, called haptic memory, lasts about two seconds in controlled work. Smell and taste lack settled durations.
  • Its capacity is large. Attention selects only a small part before the trace fades.

What sensory memory is

Sensory memory stores brief sensory events before any deeper processing. Sights, sounds, touch, smell and taste enter their own registers first. The store holds raw form, not meaning. Attention then moves selected parts to short-term memory. The rest disappears within seconds. This gateway role is the classic three-stage view of memory. OpenStax (Psychology 2e). How Memory Functions.

The store is pre-categorical. It keeps the shape of the input before the brain names it. A row of letters stays as visual form first. A spoken list stays as sound first. Meaning comes later, after attention selects the trace. This is why a location cue helps readout, while a meaning cue does not. Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.

Attention is the gate. You meet far more input than you can keep. The registers hold a wide copy for a very short time. Attention reads from that copy. What it reads can move on. What it skips fades almost at once.

The registers in one table

How long each sensory register lasts, and the study behind each number.
RegisterSenseTypical durationClassic evidence
IconicVisionAbout 250 to 500 milliseconds, usable up to about one second in Sperling's setupSperling 1960 partial report; Averbach and Coriell 1961 masking
EchoicHearingAbout two to four secondsDarwin, Turvey and Crowder 1972 auditory partial report; Crowder and Morton 1969 suffix work
HapticTouchAbout two seconds in controlled workControlled touch change work, less studied
OlfactorySmellNot settledLittle controlled work
GustatoryTasteNot settledLittle controlled work

Iconic memory is the visual register. It holds a large visual scene for a fraction of a second. A sparkler trail at night shows the idea in daily life. Each point of light lingers briefly and blends with the next. Emory University (Wyttenbach). Iconic Memory.

Echoic memory is the auditory register. It holds recent sound long enough to bind speech. When someone speaks while your attention is elsewhere, you can often replay the last few seconds. You answer after you first say you did not hear it. The words stayed in the echoic store. Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.

Haptic memory holds touch and body position for about two seconds in controlled work. It lets you sort keys from coins in a pocket by feel. Olfactory and gustatory stores exist, but their timing lacks controlled estimates. Treat them as less studied, not as equal to vision and hearing. Simply Psychology (2026). Sensory Memory.

For the closest pair in detail, read how iconic and echoic differ in detail. That article compares the two stores side by side. This article defines the whole system they belong to.

How Sperling proved the icon

George Sperling tested vision with brief letter arrays in 1960. He showed twelve letters in three rows of four for a very short flash. Then he asked observers to report them in two ways.

In whole report, observers wrote all the letters they could. They recalled about four of twelve, near one third. In partial report, a high, middle or low tone sounded right after the display. The tone told them to report only the top, middle or bottom row. They did not know the row in advance.

With no delay, they recalled about 76 percent of the cued row. That implies about nine of twelve letters were available at the end of the display. The gap between four and nine is the point. More was seen than could be reported. The limit was readout, not seeing. Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.

Sperling then delayed the tone. As delay grew, the partial-report advantage shrank. In one textbook figure the advantage is gone after about 250 milliseconds. In a university replication summary the icon is essentially gone after about one second. Answers should give the range with its setup, not one fixed number. Displays, masks and measures shift the estimate. Emory University (Wyttenbach). Iconic Memory.

A later mask can also erase the icon. Averbach and Coriell showed that a pattern on the same spot overwrites the trace before readout. A location cue helps. A category cue, such as letters against digits, does not. The store holds raw form, not meaning. Simply Psychology (2026). Sensory Memory.

What echoic memory adds

Hearing needed its own Sperling test. Darwin, Turvey and Crowder played three streams of letters and digits to the left ear, right ear and both ears at once. A visual cue after the sound told listeners which stream to report. Partial report beat whole report again. The advantage lasted longer than in vision, up to about four seconds. Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.

Two list effects point to the same store. Spoken lists show a stronger recency effect than printed lists. A spoken suffix at the end harms the last items, while a tone does not. Crowder and Morton named the store behind these effects precategorical acoustic storage. It holds the sound of recent items before meaning takes over. Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.

The store is sensitive to sound, not to meaning. A same-voice suffix harms recall more than a different-voice suffix. Lists of homophones lose the recency edge because the sounds match. Later work adds limits. How a listener parses the suffix can change the effect. Silent lip-read lists can show similar patterns. The echo is best described as acoustic-first, with stated limits, not as meaning-free in every case.

What attention does with the buffer

The buffer fills fast and fades fast. Attention decides what survives. It selects a small part for short-term memory while the trace is still there. Cuing works because it selects from the buffer before the content decays.

This also explains stable sight across eye jumps. The eyes move several times per second in rapid jumps. The icon bridges the gaps, so the world looks steady instead of lurching. Change blindness shows the other side. When a brief blank sits between two scenes, large changes can pass unseen. The buffer held the first scene only briefly, and attention never moved the change onward.

For study, the lesson is direct. Brief exposure alone leaves no lasting trace. What you attend to within seconds can move on. What you skip leaves nothing behind. Read the key detail with full attention before it fades.

Sensory memory vs short-term memory

Where the first store ends and the next store starts.
PointSensory memoryShort-term memory
RoleFirst gateway that holds raw inputWorking store that holds attended items
DurationA fraction of a second to a few secondsAbout 15 to 30 seconds
CapacityLarge, near the whole scene or streamSmall, a few items
ContentRaw form before meaningProcessed items you can rehearse
AttentionFills before selectionHolds what selection picked
LossRapid fade or maskingFade or interference without rehearsal

Short-term memory then holds attended items for about 15 to 30 seconds. OpenStax (Psychology 2e). How Memory Functions. For why depth helps study more than raw persistence, see levels of processing.

Three terms students mix up

Three pairs sound alike and test different ideas. Keep them apart with one question each.

Sensory memory against iconic memory. Sensory memory is the whole system across senses. Iconic memory is vision only. Ask which sense the item names. If it names vision, choose iconic. If it names the system, choose sensory.

Sensory memory against short-term memory. Sensory memory holds raw input for seconds or less. Short-term memory holds attended items for about 15 to 30 seconds. Ask how long the stem says the trace lasts. Seconds or less points to sensory. Tens of seconds points to short-term.

Perception against sensory memory. Perception interprets input. Sensory memory stores it briefly before interpretation. Ask whether the stem describes understanding or brief holding. Understanding points to perception. Brief holding points to sensory memory.

Students who save sensory memory, iconic memory and echoic memory to one list in Psychology World can drill these pairs with flashcards. One side names a duration or a sense. The other side names the register. Short rounds keep the numbers apart with less relearning.

Frequently asked questions

How long does sensory memory last

It depends on the sense. Vision lasts about a quarter to half a second, with a usable trace up to about one second in Sperling's setup. Hearing lasts about two to four seconds. Touch lasts about two seconds in controlled work. Smell and taste lack settled durations. Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.

What is an example of sensory memory in psychology

Seeing a sparkler trail after the light moves is an iconic example. Replaying a question you first missed is an echoic example. Sorting keys from coins by feel uses haptic memory. Each case holds raw input briefly after the stimulus ends.

What is the capacity of sensory memory

It is large. At display offset, observers hold near the whole scene or stream in raw form. Sperling's partial-report work implies about nine of twelve letters available at once. Only about four survive whole report because readout is slow and the trace fades during recall.

What did Sperling show in 1960

He showed that more is seen than can be reported. Whole report gave about four of twelve letters. Partial report with a tone right after the display gave about 76 percent of any cued row. The advantage shrank as the cue was delayed, which points to rapid decay. Emory University (Wyttenbach). Iconic Memory.

Is haptic memory real, and what about smell and taste

Haptic memory has controlled support as a brief touch store, but less than vision and hearing. Treat its two-second value as tied to controlled setups. Smell and taste stores likely exist, yet their durations lack controlled estimates. Do not state fixed numbers for them.

Sources

  1. 1.OpenStax (Psychology 2e). How Memory Functions.
  2. 2.Kahan (UMN Cognitive Psychology). Chapter 2: Sensory Memory.
  3. 3.Emory University (Wyttenbach). Iconic Memory.
  4. 4.Simply Psychology (2026). Sensory Memory.
  5. 5.Cleveland Clinic (2024). Sensory Memory.

Look the terms up yourself

18,000+ psychology and psychotherapy terms in English and German, on iPhone, iPad, Mac and Apple Vision.

See how Psychology World works
All posts