Iconic vs echoic memory: the difference, and the test that sorts any example

Iconic memory is the visual register of sensory memory: it holds what you just saw for under a second. Echoic memory is the auditory register: it holds what you just heard for a few seconds. Both are brief, high-capacity copies of raw input, and attention decides which parts move on.

This article defines each register, gives the experiment behind each duration, and ends with a sorting test you can use on any exam example.

What sensory memory is

Sensory memory is the first stage of the multi-store model of memory, as Atkinson and Shiffrin described it in 1968. Everything the senses take in lands here first, and only attended material moves on to short-term memory. It has four fixed properties. It is modality-specific, with a separate register per sense. It has high capacity, holding far more than you can ever report. It decays fast, from a fraction of a second to a few seconds. And it is pre-categorical, a raw copy of the stimulus with no meaning attached yet.

The register fills on its own, without attention. Attention then samples from it. That order matters: you do not filter the world directly, you select from a brief buffer that is already fading. Anything you do not select in time is lost, and new input overwrites it. The two registers students must know are the visual one and the auditory one.

Iconic memory

Iconic memory is the visual sensory register: a brief image of a scene that lingers just after the scene disappears. Ulric Neisser coined the term in 1967 (sensory memory account), from the Greek word for image. It holds a large amount of visual detail, and it decays within about half a second.

George Sperling showed this in 1960 with a simple but elegant design. He flashed arrays of 12 letters, in three rows of four, for about one twentieth of a second. When people reported the whole array, they named only four or five letters. That limit looked like the capacity of vision itself. Sperling then changed one thing: right after the letters vanished, a tone told the person which single row to report. Because the cue came after the display, nobody could prepare for one row in advance. With an immediate cue, people reported the cued row almost perfectly, which implies nine or more letters were available at the instant the display ended. As the cue was delayed, accuracy fell fast, and after about one second the advantage was gone. The bottleneck was never the eyes. It was the speed of reporting from a trace that fades in under a second. His paper is The information available in brief visual presentations.

Averbach and Coriell refined the timing in 1961. They showed letter rows and then a marker that named the target letter. With a bar marker, accuracy was high right after the display and fell as the delay grew, with the trace available for about 250 milliseconds. With a circle marker flashed around the target, people often could not report it at all: the later stimulus overwrote the trace in the same spot. A new visual input in the same place erases the old one before you read it out. Their work is described in the sensory memory account.

Everyday iconic memory is easy to catch. Wave a sparkler and you see a trail of light that is already gone. Glance at a clock and look away, and the face hangs for an instant. In normal vision the icon bridges the gaps caused by saccades, the rapid jumps your eyes make, so the world looks stable instead of flickering between glances.

Echoic memory

Echoic memory is the auditory sensory register: a brief echo of a sound that lingers after the sound stops. Neisser coined this term too, in 1967, by analogy with the icon (echoic memory account). It lasts longer than the icon, typically two to four seconds. The extra time makes sense because sound unfolds over time. To hear words and melody, the brain must hold earlier fragments while later ones arrive.

Darwin, Turvey and Crowder adapted Sperling's logic to hearing in 1972. They played three spoken lists at once through headphones, seeming to come from the left ear, the right ear, and the middle of the head. Listeners either recalled everything or, after a delayed cue, just one location. Partial report beat whole report again: people recalled a higher share of the cued stream than they could manage for the whole array. The advantage shrank as the cue was delayed and had disappeared by about four seconds. One detail differed from vision: location worked far better as a cue than the category of the items. Their paper is An auditory analogue of the Sperling partial report procedure.

A second line of evidence came from ordinary list recall. Heard lists show a stronger recency effect than read lists: the last items are remembered best. Crowder and Morton called the store behind this precategorical acoustic storage in 1969. Their key finding was the suffix effect. A redundant spoken word tagged onto the end of a list wipes out the recency advantage, while a non-speech tone does far less damage. The spoken suffix enters the same store and overwrites the true final items. Morton, Crowder and Prussin then showed in 1971 that the damage tracks acoustic similarity, not meaning: a suffix in the same voice and location hurts most. The store holds raw sound, not recognized words. These studies are summarized in the echoic memory account.

Everyday echoic memory is the reason you can say "what?" and then answer before anyone repeats themselves: the sentence is still in the buffer. It is also what binds a stream of syllables into words as speech arrives, and what lets a tune cohere out of separate notes.

Researchers can also measure the echo without asking anything. Mismatch negativity is an automatic brain response that appears when a rare deviant sound breaks a regular pattern, which means the brain compared the sound against a stored model of the standard (how mismatch negativity indexes the echo). Because it needs no report and no attention, it works in newborns and in patients who cannot follow instructions.

Iconic vs echoic memory: the comparison

Use one question on any example: which sense was stored, and how long did the trace last? A seen thing that survives under a second is iconic. A heard thing that survives for seconds is echoic. Duration alone can mislead near the boundary, so name the experiment too. A flashed array with a row cue is Sperling's visual test, while simultaneous spoken streams with a location cue is Darwin's auditory test.

Three exam examples show the test at work. A student sees 12 letters flashed briefly, hears a high tone, and reports the top row: iconic, because the input is visual and the cue selects from a fading image. A listener hears three word lists at once and, cued two seconds later, still reports the left-ear list: echoic, because an icon would be long gone at two seconds. A redundant spoken word at the end of a heard list erases the recency advantage: echoic, because the suffix overwrites the auditory store, which is exactly the precategorical acoustic storage effect.

What erases the trace

Two mechanisms clear the registers, and both matter for the exam. The first is overwriting by new input: the circle marker in Averbach and Coriell's study, the spoken suffix in Crowder and Morton's work. The buffer has one slot per channel, and the next arrival replaces the current contents.

The second is time itself, and a 2018 finding changed how psychologists describe it. The standard picture was gradual fading, like a photograph losing contrast. Pratte tested this by asking people to report the exact color or orientation of items after variable delays. That design separates two things older letter tasks blurred together: whether an item survives at all, and how precise the survivors are. Across four experiments, the probability of an item vanishing rose steeply with delay, while the precision of surviving items barely changed. Items do not blur. They drop out whole, a sudden death of single traces. His paper is Iconic memories die a sudden death.

One qualification belongs here. Some researchers argue the visual trace mixes two things: visible persistence, the felt lingering of brightness, and informational persistence, the readable content Sperling measured. Coltheart made this case in 1980, and his paper is Iconic memory and visible persistence. The distinction rarely changes an exam answer, but it explains why brightness manipulations and report measures can disagree.

How to use this when you study

The registers set a hard rule for learning: only attended material moves forward. Reading past a term without engaging it leaves nothing behind, because the trace decays in under a second for print and in seconds for speech. Attention is the gate, and rehearsal is what carries selected material into durable storage. Maintenance vs elaborative rehearsal shows which kind of repetition actually moves material onward, and levels of processing explains why meaning-based engagement beats rote repetition. Recall vs recognition covers the test formats your exams will use to probe what survived.

For psychology terms, the practical version is direct. When you meet a new term, say it or write it at once instead of letting the trace decay. Then save it to a list in Psychology World and drill that list with flashcards on later days. Each drill retrieves the term instead of rereading it, which is the step sensory memory alone can never do for you. Spaced repetition shows how to schedule those later sessions.