term papers categories


  social sciences


Writer’s Name Appears Here
Professor name Appears Here
Course Title Appears Here
Date Appears Here


Sperling's 1960 account of iconic memory


     The Modal model of the mind or stage theory postulates three memory components:

Sensory memory;
Short-term memory;
Long-term memory:

     In this paper we are to stress more on iconic memory, so we relate only on the basics of sensory memory. Sensory buffer or register (Sensory Information Storage) records a stimulus as a veridical representation: iconic (visual) & echoic (auditory), held for a short period of time. The sensory store has a large capacity for storing information in its original form until attention processes opt for that information which is necessary for transmission  into the short-term store.

     Visual sensory memory (Iconic): Lasts less than one second. The Sperling (1960) experiments first reveal the existence of a visual sensory register. The results support the assumptions made about the visual sensory store as having a large capacity, processing physical characteristics, and brief duration.

Click to Order a Custom Term Paper Now...

     Auditory sensory memory (Echoic): Lasts around 2 seconds. Darwin, Turvey and Crowder (1972) used a similar procedure to Sperling's experiments, but instead used auditory stimuli to investigate the nature of echoic memory. They showed that the partial report superiority effect was lost after 2 secs. inter-stimulus interval.  

Iconic memory

     Iconic memory is a visual sensory memory comprising of a large database but that lasts for very brief moments. Information on the selected channel flows through the processing system, where it is subjected to processes that result in word or object identification, the storage or retrieval of information from long-term memory, according to Broadbent's theory. But what about the information on the unattended channel? What is its fate? Is it like a telephone call to a phone in use, getting bounced back with a busy signal? Or is it like a call that gets put on hold, still available to the callee until the connection is broken?

     One deduction is that even unattended information entertains some processing; it doesn't just stop at the eye or ear. Thus, the unattended input is subjected to preattentive processes responsible for analyzing some of the basic perceptual properties of a stimulus (e.g., in vision, color and brightness). Because any process happens over time, this implies that the incoming information is stored in memory for some period of time. While it's in memory, it is subjected to preattentive analyses, and it can also be selected by the attention system for further processing. The Sperling (1960) experiments used the whole and partial report technique of post-stimulus cueing to first demonstrate the existence of a visual sensory register.

Sperling Theory 1960

     A landmark study by George Sperling (his 1960 doctoral dissertation) shed some light on the memory system, which in general is called, the sensory memory store. In the case of vision, it's often called iconic memory; in audition, echoic memory.

Click to Order a Custom Term Paper Now...

     In the initial stages of psychology, researchers were interested in something called the "perceptual span." This pursuit was to determine how much information could be gathered in a single percept. For example, when you read text, how many letters can you interpret in a single glance? Or, how many coins can you distinguish if you briefly glance at them? The goal was to identify the limits of perceptual abilities. A variety of ingenious studies seemed to draw similar conclusions that people could accurately report about 4.5 items from a brief percept.

     The above facts were built into many theories of cognition until a set of experiments by Sperling in 1960 proved it wrong. Sperling suggested that the 4.5 item limit was imposed not by the capabilities of the perceptual system, but by observers' abilities to recall items that had been seen. To test this possibility he designed a partial report experiment.

    In Sperling's original experiment, the observer saw a three by three matrix of haphazard letters on each trial. The letters were flashed for a very short period of time (50 milliseconds). After the letter matrix turned off, one of three tones sounded, and the observer reported the letters from the row associated with the tone. If the tone was presented directly after offset of the letters, Sperling found that observers were nearly 100% accurate in reporting letters from the indicated row.

     Sperling also initiated a way to gauge the duration of information in sensory store. When he increased the time between the offset of the letter matrix and the onset of the tone cue, he found that recall of letters became worse. If the tone was deferred by a full second, recall was comparable to the 4.5 items found in other experiments.

Click to Order a Custom Term Paper Now...

Tachistoscopic Identification

 The Whole Report Procedure

     Sperling's experiments were prompted by earlier investigations of the perception of tachistoscopically presented displays. A Tachistoscope is a machine that can be used to present a display that is visible for a brief interval (e.g. 10 milliseconds [ms], or thousandths of a second), and is then followed by a blank visual field. This assignment is called the "whole-report procedure" because subjects are instructed to try to report the entire display.

     Sperling thought out two other possibilities. One of these is a limitation of the response system. After the identity of a letter has been determined, its name must be kept in short-term memory (STM) until it can be reported. Given George Miller's proposal that STM has a limited capacity (in Miller's estimation, 7+/- 2 items), possibly the reason that subjects can't report more than four letters in a display is that they can't remember all of the letter names long enough to report them.

     The second possibility is a perceptual limitation. That is, perhaps the perceptual system's input channels have a restricted capacity, such that a large display exceeds this capacity. In this case, some of the information about the display would be lost before it was registered in iconic memory, and thus before it could be subjected to preattentive analysis or selected for further processing.

The Partial-Report Procedure

     Sperling considered of an ingenious method to test between these possibilities. Realizing that he needed a way of presenting a large display without requiring a response that exceeded the capacity of STM, Sperling developed the partial report procedure. As in some of his previous experiments, the displays in the partial report procedure included three rows of four letters each (and thus 12 letters per display). Rather than having the subjects attempt to report all 12 letters, however, Sperling only asked them to report the letters in one of the three rows. He did this by following the display with either a high- medium-, or low-pitch tone that signaled which of the three rows (top, medium, or bottom) to report.

     In the first experiment, the tone sounded 50 ms after the display ended, and subjects were able to report about 3.3 of the four letters in a row. Because the subjects didn't know in advance which row to report, Sperling reasoned that at the time of the tone (50 ms after the display was removed) subjects had available in iconic memory information about 10 letters (3.3 letters x 3 rows). Thus, performance in the traditional (whole-report) procedure appears to be limited by a bottleneck in the response system (i.e. the capacity of STM), and not by the capacity of iconic memory.

Click to Order a Custom Term Paper Now...

     Sperling also explored the properties of iconic memory. For example, Sperling found that performance degraded gradually as the ISI (interstimulus interval) between the display and the cue increased. This result suggests that the information in iconic memory slowly decays. However, at the limit (an ISI of one second or more), subjects still got about 4 letters correct. Sperling explained this finding by assuming that while subjects wait for the cue they make a guess about which row will be cued and start storing the names of those letters in STM. They'll guess right one-third of the time.

     In further experiments, Sperling and others showed that the duration of iconic memory depended on factors such as the brightness and contrast of the letter display and the brightness of the blank display that followed it. In addition, if a letter display is followed by another letter display (rather than a blank field), the second display seems to erase the image of the first one from iconic memory. Thus, information is lost from iconic memory due to both decay and interference (i.e. erasure by another stimulus).

Discussion for and against Sperling's 1960 account of iconic memory

     When a person looks at a scene, he seems to consciously perceive something throughout the visual field. Even at stimulus onset, conscious perception does not seem to be restricted to the object of attention. This sort of introspective argument is unsatisfying, because it is extremely difficult to test experimentally. Indeed, there are serious problems with any experimental effort to directly ask subjects if something is consciously perceived without attention. The direct approach would be to ask subjects to attend to visual stimulus "A" and report about the perception of stimulus "B". However, this proves to be impossible because the demand to report on B directs attention to B. If you do not attend to a stimulus, then you cannot report on that stimulus regardless of whether or not you saw the stimulus. Thus, the problem of perception without attention is a difficult experimental nut to crack, if one is willing to accept that it is possible to see stimuli that cannot, for one reason or another, be reported.

Click to Order a Custom Term Paper Now...

     The extensively employed solution to this methodological problem involves presenting a stimulus, arranging for the subject not to attend to that stimulus, removing that stimulus and then, after the unattended stimulus is gone, asking the subject to report on its properties. The usual finding, as seen in the examples given above, is that subjects are very poor at after-the-fact reporting on the properties of unattended stimuli. Note, however, that an inability to report is not the same as an inability to see. A subject, asked to report all the letters in a briefly flashed, 12-letter display, could report only about 25-30% of the letters. However, the subject would do much better if cued after the display was gone to report on a letter at a specific location. Sperling's subjects responded after the stimulus had been removed but subjects could not attend to the physical stimulus.

     An additional viewpoint, as in the Sperling phenomenon, is that the subjects were phenomenally conscious of all the masked letter shapes, but could not apply the letter concepts and perhaps could not apply perceptual representations required for reflexive consciousness of all of them.  Or, as before, perhaps they did briefly apply the letter concepts but with insufficient attention those conceptual representations dissolved. There is no evidence that the subjects were able to access any letter they chose.
      The common thread in the experiment has been that unattended stimuli cannot be accurately informed after they are gone. Unfortunately, there are at least two plausible explanations for this failure.

Click to Order a Custom Term Paper Now...

  1. As recommended by the inattentional blindness hypothesis, the stimuli might not be seen if they are not attended.
  2.  On the other hand, the stimuli might have been seen but not remembered.

     The result of Sperling’s theory rightfully obliterates theories that hypothesized that percepts hold only 4—5 items. Because the tone was sounded after the letters vanished, observers must be focusing on the appropriate row, as it is stored in some type of sensorystore of perceptual information. Moreover, since perfect performance is found regardless of which row is indicated, the sensory store must contain a nearly perfect representation of the total visual percept. Thus, perceptual span was not 4.5 items, but essentially every item in the visual field. What determined perceptual span in the earlier experiments was not limits of the perceptual system, but the time needed to report the seen items. The duration of information in sensory store is very brief (a few hundred milliseconds), so as observers report what they see, items in sensory store fade away. By the time observers report on 4 or 5 items, the sensory store information is gone and recall is finished.

     Tachistoscopic perception can be surprisingly good, but it is also remarkably limited. For example, when a display of 1 to 4 letters was presented for 50 ms (quite literally less than the blink of an eye), subjects could report what they had seen (ignoring the fact that sometimes one letter is misperceived as another; e.g., a C might be seen as an O).

owever, if 6, 8, or 12 letters were presented, subjects were still only able to report about four of them. Thus, there's some sort of bottleneck limiting performance here. One obvious possibility is that the limit is related to the time it takes to identify a letter. Suppose it takes about 12 ms to identify each letter. At this rate, a subject would have time to identify about four letters before the display was removed. Sperling tested this idea by presenting 12-letter displays for stimulus durations that varied from 15 to 500 ms. What he found is that stimulus duration made no difference: Subjects were able to report about as many letters (on average, just over four) from a 15 ms display as from a 500 ms display. Thus, the limitation in performance doesn't seem to be related to the processing rate for individual letters.

Click to Order a Custom Term Paper Now...


Disclaimer: These papers are to be used for research/reference purposes only. All papers should be used with proper references.


© Copyright 1996-2008 Best Term Paper and Research Papers