Psychology Of Visual Memory Research Paper

1. Visual Memory As The Channel Of Communication

In the most general textbook usage, visual memory refers to the storage of information transmitted through the visual system. According to this definition, it is a form of memory which may be compared with memory for information transmitted through other sensory systems such as auditory, tactile, or olfactory memory. The type of information that is processed (colors, music, smells) and which cannot be translated between modalities captures important characteristics of modality-specific memories. Differences between modalities in the organization of information in memory are also found. For example, when subjects are requested to remember a series of items presented in succession, in auditory presentation memory is better for the first and last items than the middle items (primacy and recency effects), whereas in visual presentation only the former effect is observed. While the basis for such modality effects is not known, they probably reflect differences in the way information is processed by the perceptual systems rather than characteristics of specific auditory and visual memory systems as such.

The channel-of-communication definition of visual memory misses the point that vision is a gate to many memory systems, and conversely that information in different sensory channels may be processed by the same amodal memory system.

2. Visual Memory As A Code

In a more restricted usage, visual memory refers to a code by which visual perceptions are inscribed and retrieved from memory in the form of images or sequences of images. According to this definition, visual memory is not a separate memory system but a representation of information that is processed by different memory systems. This definition also acknowledges that the channel of communication does not specify the memory format; what is viewed by the eye may be stored amodally, e.g., in memory systems relying on verbal codes.

2.1 Object Recognition

A direct expression of visual memory is the perceptual process itself, in the recognition of colors, textures, shapes, and objects in the world. This is achieved by some process by which on-line information from the visual system is matched to stored descriptions. Perceptual recognition is immediate, representing a fast, automatic and implicit (nonconscious) process, and it precedes semantic and episodic memory retrieval which deal with the identification of particular, familiar objects. The vast collection of memories that allow perceptual recognition is at least partly associated with a separate memory system, called the perceptual representation system (PRS) (Tulving and Schacter 1990). The PRS was formally proposed as a separate memory system in 1990, and is currently under intense investigation. It is believed to consist of several subsystems, one of which is a structural description system which stores descriptions of coherent or meaningful objects. A more recent theoretical proposal identifies a series of specialized memory processes at an earlier and more primitive level of representation, which stores information about elementary attributes of the visual world, such as the orientation, color, motion, and texture of visual patterns (Magnussen and Greenlee 1999). The function of this latter form of visual memory is not known but it might store information across fairly short time spans to help building permanent object representations.

2.2 Visual Representation In Episodic And Semantic Memory

Visual recognition goes beyond the recognition of classes of objects. The picture in the morning paper is not just a sidewalk cafe, it is the Deux Magots in Paris, and the person in the front row is not just a female figure, it is the movie star Catherine Deneuve. The conscious recognition of particulars is a visual expression of semantic memory—a memory system that stores information about general facts. And if you are a Parisian or happened to be a lucky tourist, you might also remember seeing Catherine Deneuve walking by as you were sitting outside the very same cafe, and you might remember her looks at that particular day. This production of a memory image to the ‘inner eye’ is a visual representation of episodic memory, a memory system that stores information about personal facts and episodes.

Semantic and episodic memory represent declarative or explicit forms of memory. Both personal observation and experimental research testify to the human capacity for visual episodic memory. In a series of classical studies in cognitive psychology (Standing 1973) subjects viewed large numbers of black-and-white and color slides of natural scenes, objects, persons, interiors, and landscapes, and memory was later probed by presenting pairs of slides containing one previously seen and one novel picture and asking the subjects to pick from each pair the picture they remembered. Human subjects are amazingly good at this task. With a learning set of 2,500 pictures the score was on average 90% correct after a 36 h memory interval, and with a learning set of 10,000 pictures the score was 84% correct in the test trials. This easily outperforms the memory for words tested with similar methods.

The excellent recollection of pictures and visual scenes has been explained in terms of the dual coding theory, which distinguishes between visual and verbal memory codes (Paivio 1995). Picture memory has an advantage because pictures are effectively coded in both memory systems whereas words are coded primarily, but not exclusively, in verbal systems. The effectiveness of dual coding is illustrated by memory enhancing techniques, e.g., in the ‘method of loci’ where items to be remembered are attached to places in a visualized familiar environment and later retrieved by mentally traveling the same path in memory.

The distinction between visual and verbal coding systems is also central in theory of working memory. The working memory is a system for maintaining new information or information retrieved from long-term memory, and this maintenance of information is performed by two so-called slave systems, one of which is the ‘visual–spatial sketch pad’ a hypothetical rehearsal mechanism that keeps refreshing visual and spatial perceptual information.

3. Are Visual Codes Necessary?

The necessity for postulating visual codes has been questioned by cognitive psychologists who favor a unitary memory system based on propositional, language-like codes. As an illustration, studies of memory expertise have shown that the memory of chess masters is vastly superior to that of ordinary players for real game positions, but not for random positions of chess pieces on the board (Chase and Simon 1973). The expertise memory is explained by the fact that masters have a long-term storage of a large number of visual ‘chunks,’ where a chunk is a constellation of actual game positions involving a few pieces, and the estimated number chunks is about 50,000, which is comparable to the number of words commanded by persons with a college education. Obviously, such chunks may be effectively represented either in terms of image codes or in terms of propositions—or in both.

The arguments for the existence of visual codes in memory may be summarized under four headings, as follows.

3.1 Subhuman Primates Ha E Detailed Visual Memories

Experiments with monkeys have shown that these animals can learn to discriminate between large numbers of complex abstract fractal or snow-flake type visual patterns, store these in long-term memory, and utilize the information in single-trial short-term memory tasks. Few, if any scientists have claimed elaborate propositional representation in monkeys.

3.2 High-Precision Storage Of Sensory Information

Experiments carried out in the 1990s have shown that sensory information, i.e., texture, orientation, motion, and similar attributes of the visual image, is extremely well retained in short-term memory. In fact, attributes such as motion and spatial frequency are represented in memory with the fidelity of the sensory image, at least for memory intervals of the order of minutes, but there is also some evidence for long-term storage of detailed sensory information. In one famous experiment, an eidetic subject was able to identify the three-dimensional figure hidden in random-dot stereograms when the left and right-eye patterns were viewed on different days (Stromeyer and Psotka 1970). Such high-fidelity images cannot reasonably be stored in terms of propositions.

Long-term memory changes in the form of perceptual learning are also well established. Subjects get better at performing fine discriminations between orientations or textures across large number of training sessions. The learning is specific for the training conditions and does not transfer to other conditions; for example, when subjects are trained in discriminating visual textures in one part of the visual field, the training effect does not transfer to other parts of the visual field (Karni and Sagi 1991). Some forms of visual memory must be due to local neural changes closely associated with the perceptual process itself.

3.3 Dissociation Between Visual Memory And Other Forms Of Memory

Lesions of the brain which disturb semantic and episodic memory may leave the perceptual recognition process intact. The patient’s perception of the world is normal but he or she does not remember seeing these objects or persons before. Recognition memory may be demonstrated though perceptual priming: If the patient is shown a drawing of an object and at a later time is shown the same drawing again, there is a gain in the time required to identify and name drawings that were previously seen compared with novel drawings, and the gain is comparable for amnesic and normal subjects.

If, on the other hand, the link between visual memory representations and on-line perception is disrupted, amodal (verbal) episodic and semantic memory may be intact but the person can no longer recognize what he or she sees. In this condition of visual agnosia there may be nothing wrong with perceptual process per se, the person navigates safely in the geographic environment and may even copy diagrams and drawings, but the information which comes from vision is meaningless.

3.4 Visual Processing Areas Of The Brain Are Recruited In Memory Tasks

Studies of memory processes in the human brain with modern brain imaging technology such as positron emission tomography (PET) and functional magnetic resonance tomography (fMRI) have shown that there is no special region of the brain where memories are stored. Memory involves widely distributed networks and all major regions of the cerebral cortex have been implicated in the various forms of memory. Since visual memory is a form of representation, a code rather than a system, one would expect it to share processing space with several memory systems, in addition to regions of the occipital lobe that are specific for the processing of visual signals.

Memory researchers now generally agree that memories may involve many of the neural ensembles that were involved in the processing of the information in the first place. In vision a close link between perception and memory is illustrated by the observation that certain forms of perceptual deficits due to local brain lesions appear to affect memory as well. For example, persons who lose their ability to see colors owing to bilateral lesions of a region of the occipital lobe known as V4 usually also lose their ability to remember or imagine colors.

The current debate on the neural mechanisms of visual memory does not consider whether sensory processing areas are recruited in memory, but how far down the processing hierarchy recruitment extends. Some brain imaging studies claim that the primary visual cortex (region V1), i.e., the very earliest level of sensory analysis receiving direct input from the eye, is activated in visual imagery tasks, but other studies do not find this activity (Cabeza and Nyberg 1997).

A logical argument may be directed against the involvement of early visual processing areas and a unification of visual perception and memory processes: Since memory images are normally much less vivid and detailed than perceptual images, and memory is rarely confused with perception, they must activate at least partly different neural circuits. The precise relationship between visual perception and visual memory is a major task for future research.

Bibliography:

1. Cabeza R, Nyberg L 1997 Imaging cognition: an empirical review of PET studies with normal subjects. Journal of Cognitive Neuroscience 9: 1–26
2. Chase W G, Simon H A 1973 The mind’s eye in chess. In: Chase W G (ed.) Visual Information Processing. Academic Press, New York, pp. 215–81
3. Farah M 1990 Visual Agnosia. MIT Press, Cambridge, MA
4. Karni A, Sagi D 1991 Where practice makes perfect in texture discrimination: evidence for primary visual cortex plasticity. Proceedings of the National Academy of Sciences of the USA 88: 4966–70
5. Magnussen S, Greenlee M 1999 The psychophysics of visual memory. Psychological Research 62: 81–92
6. Paivio A 1995 Imagery and memory. In: Gazzaniga M S (ed.) The Cognitive Neurosciences. MIT Press, Cambridge, MA, pp. 977–86
7. Standing L 1973 Learning 10,000 pictures. Quarterly Journal of Experimental Psychology 5: 207–22
8. Stromeyer C F, Psotka J 1970 The detailed texture of eidetic images. Nature 225: 346–49
9. Squire L R, Kandel E R 1999 Memory. From Mind to Molecules. W. H. Freeman, New York
10. Tulving E, Schacter D L 1990 Priming and human memory systems. Science 247: 301–6
11. Zeki S 1993 A Vision of the Brain. Blackwell, Oxford