What is sensory memory?

Introduction

Human senses—such as sight, smell, touch, and hearing—pick up information about the surrounding physical world. Once this information is received, it must be converted by the senses into a code that is transmitted to the brain and eventually interpreted. Sensory memory plays a critical role in this process of transforming the outside world into an inner psychological experience. The sensory-memory system stores information acquired through the senses for a brief time to allow other memory systems to screen and select which parts of the message will be kept for further processing. Depending on the particular sense modality, the duration that items can be held ranges from 0.25 to 2 seconds. Thus, one primary characteristic of this system is that its retention is very brief.

One important question concerns how much information can fit into sensory memory. Sensory memory has a larger capacity than short-term memory, which can hold five to nine bits of information, but a smaller capacity than long-term memory, which is limitless. A third, and somewhat controversial characteristic, of sensory memory is that the information it holds is believed to be “precategorical.” That is, the information has not been significantly altered, categorized, or processed but is believed to be represented in a form that is nearly identical to its original copy. However, research by Phil Merikle published in 1980 weakened this third distinction of sensory memory from other memory systems.

Merikle performed one experiment in which he demonstrated that subjects could pick out a group of letters from any array of letters and numbers that were held in sensory memory. Merikle’s research provided evidence that information in sensory memory is susceptible to at least some processing. This assumes that the test stimuli were held in sensory memory and never transferred into short-term memory. If the information had been transferred, this second memory system could have been responsible for the ability to select out a group of letters from a group of letters and numbers.

Each sense modality is believed to have its own distinct sensory memory with its own unique characteristics. Much more is known about visual and auditory sensory memory than any of the other senses, simply because the vast majority of research has focused on these two systems. Ulric Neisser coined the terms “ iconic memory” and “ echoic memory” to refer to these separate systems.

Iconic Memory

Much of what is known about iconic memory has come from an experimental procedure originally used by George Sperling in 1959. He wanted to find out how much information could be seen during a rapid, brief exposure of stimuli. To answer this question he presented subjects, using specialized equipment, with a matrix of twelve letters (four letters placed in three rows) for the duration of 50 milliseconds (0.05 seconds). After asking the subjects to recall as many of the letters as they could, Sperling found that, on average, they named about four items. More important, he noticed that his subjects insisted that they had seen more than four items but had forgotten the others. Sperling concluded that perhaps more of the letters were originally seen but the subjects held them for such a brief time that some of the items were lost through memory decay as the subjects called out the letters.

To explore this possibility further, Sperling altered his experimental procedure. Rather than have subjects call out items from anywhere within the matrix of letters—called the whole-report technique—he presented the test subjects with a tone signal that immediately followed the presentation of letters. He used the tone to cue the subjects to call out only items from one row of the matrix. For example, a high tone would signal that the top row should be recalled. Using this partial-report technique, Sperling found that the subjects recalled three to four letters, despite the fact that they did not know in advance which row they would be asked to report. This level of performance was nearly identical to the previously used whole-report technique and led Sperling to infer that the subjects saw approximately nine items. He based this conclusion on the fact that three items were recalled from a row that was not determined until after the display was shown. Since it did not matter which row was signaled for recall, Sperling found that, on average, three letters were seen from each of the three rows. Subjects did not call out all nine letters because of the short duration of the memory trace.

Sperling’s research provided empirical support for the existence of a brief sensory register. Not only did he find the capacity of sensory memory, but also, by delaying the presentation of the tone during the partial-report technique, he learned about its duration. He found that information in sensory iconic memory lasted only about 250 milliseconds (0.25 seconds). Sperling published his findings in the 1960 Psychological Monographs article, “The Information Available in Brief Visual Presentations.”

By the late twentieth century, many researchers referred to iconic memory as visual sensory memory and established two types of its persisting visual information, known as visible persistence and informational persistence. In visible persistence, neurons and photoreceptors influence how images briefly remain then fade, somewhat like what a person sees when a flashbulb illuminates a dark room. In contrast, informational persistence refers to information that remains available and does not fade after stimuli occur. The partial-report method developed by Sperling relies on this type of visual sensory memory, which researchers further differentiate into two types according to information decay, organization, and categorical factors.

Echoic Memory and Other Senses

Echoic memory has been ingeniously studied using a similar method as iconic memory tests. Subjects are aurally presented with different letter combinations simultaneously to the right ear, left ear, and both ears (for a total of nine items) using special headphones. Subjects then receive visual cues to recall only the items presented to one ear. When this partial-report technique was compared to the whole-report method, the results paralleled what was found with iconic memory. Subjects recalled more items using the partial-report technique, which indicated an echoic memory was present. Most believe that the duration of echoic memory is considerably longer than iconic memory, somewhere between 2 and 3 seconds.

By the early twenty-first century, researchers concentrated on specific auditory sensory memory investigations. Elisabeth Glass, Steffi Sachse, and Waldemar von Suchodoletz emphasized the importance of young children’s ability to store auditory information in short-term memory for cognitive development associated with learning and language problems. In 2005, French scientists Julien Besle, Alexandra Fort, and Marie-Hélène Giard published findings in Experimental Brain Research, hypothesizing that auditory sensory memory sometimes interacts with visual information. A 2006 Journal of Cognitive Neuroscience article addressed timbre and auditory sensory memory.

Researchers also investigated the role of touch in sensory memory. Some experiments that tested subjects with tactile stimuli resembled those conducted for iconic and echoic memory. Displays contained from one to six stimuli that were touched simultaneously against subjects’ arms or legs, and subjects stated how many stimuli they felt or commented about stimuli positions on the display. Textures of food and drink influence memory more than tastes. In tests, participants fasted prior to eating a breakfast consisting of yogurt, biscuits, and a drink. By nighttime, they were presented those same food types but prepared in five different textures, to test their memory to remember which version they had consumed. Subjects did not recall foods because they liked their taste but because the foods were creamy or crispy.

The role of smell in sensory memory has received less attention than other senses from researchers. Frank R. Schab and Robert G. Crowder edited Memory for Odors (1995), the first monograph focusing on that subject. They criticized much of the existing olfactory memory research as too narrow and for not incorporating sufficient memory theory. Schab and Crowder urged scientists to consider how odors were stimuli to store and retrieve memories and to differentiate how smell varied from other sensory-memory types. A 2006 Neural Computation article reported scientists’ observations of neurons in honeybees’ brains when antennal lobes, similar to olfactory bulbs in vertebrates, were exposed to odor stimuli. The study attempted to comprehend how olfactory sensory memory functioned.

An important issue that has received some attention concerns the reasons why sensory memory is necessary. In 1989, Margaret Matlin pointed out that humans live in a world in which their senses are overwhelmed with stimulation that can change in an instant. A memory system that can rapidly absorb information from this abundance of varied sensory stimuli, and retain it for even a brief time, allows subsequent memory systems to perform more in-depth analyses. In this way, information deemed to be unimportant can be discarded, while information regarded as important can advance to the next processing stage, short-term memory.

Uses of Sensory Memory

Sensory memory plays an integral role in a variety of psychological processes. For example, someone who has just returned from Europe might show a friend some of the pictures she has taken. One of these pictures shows her in Rome standing next to a large piece of stone called Trajan’s column. The process of pattern recognition gives the friend the ability to recognize her as well as the other objects in the picture. For this to occur, the visual system must receive information from the picture by attending to it; then, this information is temporarily stored in iconic memory. At this point, an interaction occurs between information held in sensory memory and memories of previous experience held in long-term memory. As mentioned earlier, information in iconic memory must be acted upon quickly or else it will be lost. Decision rules must be applied at this stage to determine which pieces of information are relevant and deserving of more complex processing. One way meaningful information can be identified occurs through a process of matching information stored in long-term memory with the contents of iconic memory. One friend can recognize the other in the picture because he “knows” what she looks like. Specifically, he has detected unique features, such as her eyes or the shape of her nose, that tell him for certain that the person in the picture is his friend. Pattern recognition does not end with identifying the friend, but extends to all the other objects in the picture. Although numerous theories have been put forth to explain pattern perception, a common component of many of these theories is the important role sensory memory plays.

In 1981, the notion of a sensory memory was used by Robert Solman and colleagues to explain the phenomenon of the word superiority effect: People who are briefly presented with a letter (T) and then asked to choose which letter they have just seen from a set of two letters (D, T) do not perform as well as when they are presented with a word (CART) and have to choose the correct word from a set (CARD, CART). Basically, subjects are better able to discriminate the letter T from D when it is embedded in a word. Solman and colleagues believe that one possible explanation for the word superiority effect is that words are stored longer than individual letters in sensory memory. Thus, the longer the words are retained in iconic memory, the more easily subjects can remember the correct response in a subsequent recall test.

Henry Ellis and Reed Hunt, in their book Fundamentals of Human Memory and Cognition (4th ed., 1989), describe how research into sensory memory can be used to help solve an important problem associated with a learning disability. They mention a particular reading problem called specific reading disability, which prevents children from reading normally. One unusual characteristic of this reading disorder is that in its early stages, no other intellectual abilities are affected. Once a child matures to the age of approximately twelve years, academic performance in other areas such as mathematics begins to deteriorate. It is not known what causes the specific reading disability.

Ellis and Hunt mention that some believed that children with the reading problem did not see the same perceptual images as normal readers. The perceptual deficit hypothesis theory assumes the root of this problem to be perceptual in nature. One unfortunate aspect of the perceptual deficit hypothesis was that it implied the problem was occurring at the most basic level of the visual system: the sense receptors. If this were true, then no amount of training, practice, or learning strategies could possibly help overcome the disability.

In 1977, Frederick Morrison and his colleagues performed an experiment using a group of sixth-graders to test this hypothesis. Half the children in the group were normal readers, while the other half was poor readers with reading disabilities. Morrison believed that if the problem were caused by a perceptual abnormality, it would become evident after looking at the precise nature of the information held in the sensory memory. By conducting an experiment similar to the one performed by Sperling, Morrison tested the contents of the perceptual stimuli in iconic memory for both the normal and poor readers to find out if they differed.

Morrison found that good readers and poor readers performed equally well on a recognition task for information in sensory memory. This finding was contrary to the prediction made by the perceptual deficit hypothesis. Morrison found that good readers began to outperform poor readers only after the cue that started the recognition test was delayed by at least 300 milliseconds. This revealed that the problem that caused the reading disability was occurring at a higher level of information processing rather than at the sensory-memory stage.

Investigation History

The notion that humans possess a brief visual sensory storage was first proposed by Wilhelm Wundt as early as 1899. Although he did not have access to the sophisticated equipment needed to demonstrate experimentally the existence of this memory system, perhaps he saw evidence for it by making the following observation: If a hand-held candle is moved rapidly in a circular path in a darkened room, one can see what appears to be a brightly lit ring. Despite the fact that the flame of the candle does not occupy every position on this circle continuously, an illusion occurs whereby one can still “see” the flame after it has moved. This phenomenon occurs as a result of visual sensory memory. Possibly, Wundt proposed a brief visual register after seeing evidence for it with common, everyday examples such as this.

In 1958, Donald Broadbent proposed a theory of attention that incorporated the concept of a sensory register. The sensory register, as Broadbent saw it, was a temporary short-term store for information received by the senses. Information from different senses was transmitted on separate channels. A selective filter was believed to act on the information in the sensory store by conducting a rudimentary analysis of its contents to determine which information should be processed more thoroughly. The notion of a brief sensory store was an integral part of Broadbent’s theory, as well as of other theories of attention which were developed later.

In 1960, when Sperling discussed his results, which provided empirical support for a sensory memory, the scientific community readily embraced them. By 1968, Richard Atkinson and Richard Shiffrin had incorporated sensory memory into their information-processing theory of memory. The Atkinson and Shiffrin theory evolved into one of the most significant and influential models of human memory devised to that time. Sensory memory, according to their model, was viewed as the first stage in a series of information-processing stages. Sensory memory captured and briefly held information while control processes determined which information would be transferred to short-term memory and eventually to long-term memory.

Although at times the concept of a sensory memory has been challenged, it continued to be a pivotal psychological construct for the understanding of many perceptual and learning processes. In the late twentieth century and early twenty-first century, psychologists initiated new sensory-memory research or returned to previous investigations (to determine if results retained credibility decades later), altered methods, or utilized advanced computerized and testing technologies. Despite scientific journals publishing these innovative sensory-memory findings, many textbooks presented outdated and inaccurate information which simplified sensory memory instead of discussing its complexities.

Sensory memory researchers explored and presented novel insights and suggested theories for future investigations. Many researchers sought to comprehend neural activity associated with sensory memory. Several experiments contemplated the effect of blinking on iconic memory. Referring to predecessors’ efforts in this field, Mark W. Becker and University of California, San Diego Department of Psychology colleagues tested subjects’ visual sensory memory in terms of noticing changes in arrays. The researchers reported in 2000 that that they had determined that a blank interstimulus interval (ISI) placed between the initial array viewed and an altered array lowered participants’ detection of changes. If the ISI contained a cue identifying where the change occurred, more subjects noticed the change.

Laura E. Thomas and David E. Irwin, visual sensory-memory experts at the University of Illinois, also studied how blinking affected cognitive processes (2006). In a partial-report iconic-memory test, they displayed arrays with letters for subjects to view during specified durations measured by milliseconds. Thomas and Irwin eliminated other factors, such as lighting fluctuations, to determine blinking was the main impediment affecting how subjects’ iconic memory functioned. They introduced the concept of cognitive blink suppression, referring to how blinking interfered with cognition. Thomas and Irwin acknowledged that neural disruptions and reduced area VI activity in brains might also disrupt iconic memory.

Christian Keysers and his colleagues studied neurophysiological aspects of iconic memory (2005). They arranged viewings of rapid serial visual presentation (RSVP) images, using ISI at lengths as long as 93 milliseconds in some RSVP tests and as short as zero in others. Test subjects included humans and macaque monkeys. During ISI, neurons in the temporal cortex of the subjects’ brains kept responding to stimuli they experienced before ISI occurred despite the absence of stimuli in ISI. As a result, the researchers hypothesized that temporal cortex neurons were important to iconic memory.

In addition to iconic-memory studies, researchers considered other factors that might affect sensory memory, including aging, Huntington’s and Alzheimer’s diseases, genetics, alcoholism, and affect. Some investigators studied how blindness and deafness affected people’s sensory memory. Researchers at the University of California, Irvine Center for the Neurobiology of Learning and Memory examined the relationship of stress hormones and memory. Subjects with post-traumatic stress disorder presented interesting research possibilities because sensory experiences associated with what people saw, heard, touched, or smelled when traumatized shaped their memories.

Bibliography

Becker, Mark W., Harold Pashler, and Stuart M. Anstis. “The Role of Iconic Memory in Change-Detection Tasks.” Perception 29 (2000): 273–286. Print.

Caclin, Anne, et al. “Separate Neural Processing of Timbre Dimensions in Auditory Sensory Memory.” Journal of Cognitive Neuroscience 18.12 (2006): 1959–1972. Print.

Galán, Roberto F., et al. “Sensory Memory for Odors Is Encoded in Spontaneous Correlated Activity Between Olfactory Glomeruli.” Neural Computation 18.1 (2006): 10–25. Print.

Gallace, Alberto, Hong Z. Tan, Patrick Haggard, and Charles Spence. “Short Term Memory for Tactile Stimuli.” Brain Research 1190 (2008): 132–142. Print.

Glass, Elisabeth, Steffi Sachse, and Waldemar von Suchodoletz. “Development of Auditory Sensory Memory from Two to Six Years: An MMN Study.” Journal of Neural Transmission 115.8 (2008): 1435–1463. Print.

Keysers, Christian, Dengke K. Xiao, Peter Földiák, and David I. Perrett. “Out of Sight but Not Out of Mind: The Neurophysiology of Iconic Memory in the Superior Temporal Sulcus.” Cognitive Neuropsychology 22.3/4 (2005): 316–332. Print.

Kuhbander, Christof, Stephanie Lichtenfeld, and Reinhard Pekrun. "Always Look on the Broad Side of Life: Happiness Increases the Breadth of Sensory Memory." Emotion 11.4 (2011): 958–964. Print.

Luck, Steven J., and Andrew Hollingsworth, eds. Visual Memory. New York: Oxford UP, 2008. Print.

Matlin, Margaret W. Cognition. 8th ed. Hoboken: Wiley, 2013. Print.

Mojet, Jos, and Egon Peter Köster. “Sensory Memory and Food Texture.” Food Quality and Preference 16.3 (2005): 251–266. Print.

Spachtholz, Philipp, Christof Kuhbandner, and Reinhard Pekrun. "Negative Affect Improves the Quality of Memories: Trading Capacity for Precision in Sensory and Working Memory." Journal of Experimental Psychology: General (14 Apr. 2014): n.pag. Digital file.

Thomas, Laura E., and David E. Irwin. “Voluntary Eyeblinks Disrupt Iconic Memory.” Perception & Psychophysics 68.3 (2006): 475–488. Print.

Vlassova, Alexandra, and Joel Pearson. "Look before You Leap: Sensory Memory Improves Decision Making." Psychological Science 24.9 (2013): 1635. Print.