A Four-Stage Hierarchical Model of Image Construction and Drawing Production:
Evidence from Visual Hallucinations, Development and Pathologic Regression in Art
Paul C. Vitz
New York University and The Institute for the Psychological Sciences
New York University
Vitz, P. C., & Kamorina, T. (2014). A four-stage hierarchical model of image construction and drawing production: Evidence from visual hallucinations, development and pathalogic regression in art. International Journal of Neuropsychotherapy, 2(1), 2-26. doi: 10.12744/ijnpt.2014.0002-0026
A four stage hierarchical model of image construction is proposed. Stage 1 involves amorphous or scribble fields of light, color and spots; Stage 2 involves abstract geometric shapes; Stage 3 schematic recognizable objects; Stage 4 represents three-dimensional realistic images. A similar four stage model is also proposed for drawing an image. Evidence to support the model of the visual construction of an image is based on phosphenes and various other kinds of visual hallucinations. Evidence supporting a similar 4-stage model for drawing an image is based on the development of children’s drawings and from drawings of adult artists suffering from mental pathologies or from Alzheimer’s disease. In addition findings from brain research supporting both models and implying underlying neurological functions and locations are cited.
Images: All images referenced in this article can be found in the PDF version – please download the PDF from the button below
In this paper we propose that a perceptual image is normally processed, that is “constructed”, in the human visual system through what can be considered four qualitatively different hierarchical stages. We also propose that the same four stages are involved in the physical production of an image, such as drawing a picture. After providing evidence for multi-stage hierarchical visual construction, we relate these stages to artistic development in children, and to changes in the depictive style of artists suffering from neurological and/or psychiatric pathology. The model involves hierarchical processing from a) initial sensing of light and color in amorphous or point fields to b) processing of perceptual representations of geometric shapes to c) cognitive schematic construction of objects to d) a normal synthesized 3-dimensional realistic representation. This hierarchical model is consistent with neurophysiological findings that suggest serial processing in ascending the neuraxis (Bar, 2003, Hawkins, 2004, Jackendoff, 1987, Prinz, 2000), yet also assumes a parallel form of processing within each functional ‘level’ of the hierarchy.
It is important to note that this is not a model of how the perception of color/light, or geometric shapes, or schematic objects or complete scenes themselves presumably take place. Instead, these stages are assumed to take place, probably using some kind of parallel processing, but otherwise how they take place is not treated. Thus this paper outlines a gross structural model of the stages or processes and the implications of such a hierarchical structure for
how an image is constructed in the final perceptual representation and how this relates to such phenomena as children’s drawings, visual hallucination, and regression in drawing by people suffering from serious pathology such as Alzheimer’s disease.
One consequence of the model is to show how a patient’s art can be directly relevant to medical practice since the model proposes that for certain patients their art is an intelligible expression of underlying neurophysiology.
The Model: Perceiving or Constructing an Image
Stage 1. The first stage or level we term sensory and it is characterized by amorphous color fields and scribbles. The visual appearance or phenomenology of Stage 1 consists of a field of light or color. Also included are spots or dots of light that occur in a haphazard pattern, that is, the dots are not arranged in such a way as to allow straight lines or other recognizable shapes. Scribbles – that is, patterns made up of many heterogeneous lines without any recognizable shapes or objects are also representative of Stage 1. Such scribble patterns are interpreted as another form of amorphous or inarticulate visual field. Scribbles are commonly made by children using a relatively sharp pointed pencil or crayon, however, if they had been given a relatively wide brush their “scribbles’ would look much like an amorphous color field. Familiar examples from modern art of images whose visible aspects are assumed to have been primarily completed in Stage 1 include the typical color field paintings of Mark Rothko, the familiar drip paintings of Jackson Pollock, and many of the works of Cy Twombly, Franz Klein and many abstract expressionists. Similar images also at Stage 1 are produced by children when they first begin drawing and are commonly known as scribble drawings.
To describe a painting as at Stage 1 (i.e. sensory: amorphous or color field; or scribble) in no way criticizes the art work or implicitly equates it with the work of a child. On the contrary, the greater aesthetic and cultural significance of the artist’s work based on higher order processing is assumed. We only illustrate image types – nothing more. (Also note when we use the expression “Stage 1, 2, 3 or 4 image” we are using shorthand to refer to an image where the visual processing of the stimulus, according to the model, has been essentially completed at the designated Stage.)
When we characterize the processing of a visual stimulus as being completed at a given stage we do not mean to imply that the stimulus is not also processed at higher levels. This is especially true of paintings by adult artists. Rather, we posit that that the stimulus is not processed any further in terms of the image properties characteristic of the stage in question. It is well known that network processing that engages the cortex brings lower visual material into a complex framework involving higher order visual associations, in addition to neural networks that engage memory, recognition, interpretations, meaning, and other contextual information. We also assume (See Zeki, 1998) that the stage represents both a level of completed processing and a level of perception.
Stage 2. The second stage is called “perceptual”, and is characterized by geometric lines and patterns (e.g. triangles, ellipses, composites of such shapes, etc.), but without any recognizable objects. These shapes we assume involve some prior processing at Stage 1. Examples of such Stage 2-processed images are the typical works of such artists as Victor Vasarely, many by Wassily Kandinsky, Piet Mondrian, Kazimir Malevich, Josef Albers and countless more recent artists whose paintings are of simple or complex abstract geometric shapes.
Stage 3. This stage consists of cognitive-schematic (i.e. cartoon-like) objects. Stage 3 is distinguished from Stage 2 by the clear appearance of recognizable objects; however these are schematic, and generally flat – such that major cues to depth (e.g. shading, linear perspective) are absent. We believe this level to be legitimate and important, although it has not been given much attention in the experimental literature. Examples of Stage 3 images include many works by Pablo Picasso, Fernand Leger and other cubists. Marcel Duchamp’s “Nude descending a staircase” is another well known representative of this stage. Of course, many artists have done a few works in this style. A recent Stage 3 artist would be Keith Haring and, of course, almost all cartoons and characitures represent what we are calling Stage 3.
Stage 4. This, the model’s final stage of perceptual construction is characterized by comprehensive and realistic 3-dimensional scenes. Stage 4 consists of the images characteristic of normal visual experience. While there are additional stages and processing beyond Stage 4, we maintain that these involve post- or extra- visual neural functions such as language and judgments about beauty for example, as identified by Kawabata & Zeki, 2004. These are beyond the scope of the present discussion. Examples of Stage 4 images include all typical realist portraits, landscapes and still life’s.
It is important to note that this model proposes that there are only four qualitatively different kinds of visual image – one type at each stage. We assume that most natural three-dimensional visual experience involves the earlier constructive use of the three “simpler” stages. A few natural stimuli are presumed to cease being perceptually processed at early stages, for example perception of the sky and the ocean are normally finished by Stage 1 although the moon and sun, when present, bring in some of Stage 2. In terms of the Model the perceptual construction of many human artifacts such as art, buildings, cartoons, etc. are often completed by Stage 2 or 3.
In addition we understand that the stages of image processing are mental, (i.e. perceptual, experiential and/or phenomenological) in character (See Zeki, 1998) and most likely reflect complex activity of brain networks that engage hierarchical logic rather than just a specific localization in the visual neuraxis. Nevertheless, we posit that there is plausible neurological evidence for some stage localization as presented below.
Perhaps the simplest evidence for the first two stages can be observed when gentle pressure is applied to the closed eyes (Oster, 1970; Tyler, 1978). When this occurs, it is common to first observe changes in the brightness of what was an all black field, followed by points of light (often colored) on a black or almost black field, and finally grids or gratings or fields of squares, that are often colored. Presumably pressure applied to the front of the eyes initially stimulates retinal elements, then the lateral geniculate, and then areas of the primary visual cortex. After this level, visual stimulation achieved by pressure presumably is no longer effective and dissipates. (For the history of study and speculation about such eye deformation-induced phosphenes see Grosser and Hagner, 1990). In other words, simple activation of the initial transductive and conductive elements in the visual pathways, without photic input, is sufficient to generate early stage responses. But the absence of additional stimuli (and differential engagement of amplifying and rectifying elements in the retina) means pressure on the eye is insufficient to elicit higher stage processing activity.
Electrically induced phosphenes
The most systematic investigation of phosphenes has involved the use of mild electric current. (Höfer, O., 1963; Knoll, M., Kugler, J., Eichmeier, J. & Höfer, O., 1962; Knoll, M., Höfer, O. & Kugler, J., 1966; Knoll, M. & Welpe, E., 1968; Seidel, D., Knoll, M. & Eichmeier, J., 1968). In these experiments, the participant was blindfolded and low level electrical stimulation was applied to the lateral orbit. Such electrical stimulation invariably elicited a visual response (i.e.- electrical phosphenes) in all observers who were then instructed to draw examples of their visual experiences. Analyses of hundreds of such reports were classified into 15 fundamental types, as shown in Figure 1. We refer to these as the Munich taxonomy of phosphenes (noting that the studies were conducted in Munich, Germany).
With the exception of spots, all of the reported phosphenes induced by electrical stimulation can be categorized as Stage 2 shapes in accordance with the present model. The intensity and frequency of electrical stimulation that elicited a given phosphene shape was not consistent across subjects, although it was consistent within a subject. The similarity of these basic shapes to most children’s basic scribbles (see Figure 3) was noted by Kellogg, Knoll and Kugler (1965).
Some of the earliest descriptions of perceptual experience indicative of the underlying visual stages can be found in the sensory deprivation literature. In these studies (male) college students spent time in a cubicle in which sensory experience was markedly reduced. The participants wore translucent goggles so that all visual input appeared as a relatively homogeneous field of light. They also wore gloves and cardboard cuffs; the latter extended from the elbow to below the fingertips, in order to reduce tactile stimulation. Auditory stimulation was limited by continuous background noise created by fans and air conditioners and by a U-shaped rubber head pillow that acted as a sound baffle. Participants lay on a bed in a cubicle for 24 hours a day, with interruptions only to eat and for toilet breaks. Most could tolerate sensory deprivation for only two or three days before finding the experience unpleasant, and terminating involvement in the experiment.
Twenty five of the 29 participants reported some form of visual hallucination. Heron (1961) stated that hallucinations typically progressed from simple to more complex: “… the first symptom would be a lightening of the visual field. Dots of light or lines would then occur. Next … geometrical figures and patterns, generally composed of reduplicated figures, followed by isolated objects against a homogeneous background. Finally, full blown scenes would appear”. The reduplicated geometrical figures were sometimes described as “wallpaper patterns”. Actual figures or objects without a background included images of “… a row of little yellow men with black hats on and their mouths open”, “… a German helmet”, and “… a procession of eyeglasses marching down the street”. The figures were often described as cartoonish. Complete scenes and landscapes were the least frequently reported hallucinations.
We believe that these reports describe all four stages or levels of visual processing as proposed. The lightening of the visual field and points of light (i.e. earliest experiences) show Stage 1. In pressure phosphenes these experiences often involve desaturated colors, although color was not addressed or reported in these experiments. Sensory deprivation-induced hallucinations at Stage 2 are represented by reports of lines and geometrical figures and patterns (e.g. wallpaper patterns). Stage 3 visual phenomena include cartoon-like figures (e.g. “… rows of little yellow men with black hats and their mouths open”), and Stage 4 is represented by more integrated scenes, especially those described “… as landscapes”. In addition, it should be noted that these perceptions occurred in the predicted order- i.e. from simple to more complex- and each higher stage was less frequently reported than the previous stage, with the highest stages reported by only about 20% of participants. (The results cited above are summarized from the work of Bexton, Heron & Scott, 1954; Heron, 1957; and Heron, 1961).
Vernon and McGill (1962) also present a multi-level description of sensory-deprivation-induced hallucinations. However, the researchers limit these effects to a 3-level process: Level 1 consisted of flashes of dim glowing or flickering light without shape. Level 2 hallucinations were more complex and were geometric, e.g. squares, circles, lattice-work; level 3 were complex and contained integrated and/or animated scenes that were highly structured and rich in detail (p. 149). While levels 1 and 2 are similar to our model’s first 2 stages, Vernon and McGill group into a single category (their level 3) what we distinguish as Stage 3 and Stage 4 processing.
A well known medical phenomenon involving visual hallucinations is the Charles Bonnet Syndrome. This syndrome describes the spontaneous appearance of complex visual hallucinations in adults who are without psychopathology. These hallucinations are experienced primarily by older patients who have visual deficits, especially at the retinal level. Reduction in vision due to vascular pathology in the periphery as well as in visual areas of the brain (e.g. A17-19) is associated with the syndrome. Schultz and Melzack (1991) noted that there is considerable literature indicating that the loss of visual input to the brain as a result of pathology can be a cause of visual hallucinations, and suggest that these “… represent a visual analogue of the phantom limb phenomenon – that is, experience generated by brain activity in the absence of sensory input” (Schultz & Melzack, 1991, p. 809). Cole (1992) has noted the similarity of sensory deprivation-induced hallucinations to those of Bonnet syndrome: he considers that absence of visual stimulation in Bonnet’s patients is analogous to the absence of external stimuli in sensory deprivation.
Bonnet syndrome hallucinations are “… usually colorful, well-formed images that are detailed and sharply in focus even if they appear at a distance” (Schultz & Melzack, 1991, p. 810). The objects seen are quite varied, with images of people and animals being especially common. Examples include “… a brightly colored circus troupe”, “… large chickens wearing shoes”, etc. (Schultz & Melzack, 1991, p. 811).
A review of the different images, as provided by Schultz, Needham, Taylor, Shindell, & Melzack (1996), reveals that 23% of subjects reported simple light flashes, colored light, black dots or “bugs”. These reports describe what we categorize as Stage 1 images. Twelve (12) percent reported formed geometrical shapes or elaborate designs, what we consider to be Stage 2. Schultz (August 16, 2000, Personal Communication) reported that Bonnet’s patients frequently described some of their images to be “…like cut-outs” and stated that the term “schematic” seemed appropriate for such descriptions. Thus, we conclude that many of the complex hallucinations reported by Bonnet’s patients are what we classify as images characteristic of Stage 3: mainly flat, schematic, and cartoon-like.
Further evidence for Stage 3 images can be derived from descriptions of Bonnet’s Syndrome that appear in the earlier French literature. As discussed by Flournoy (1902), Charles Bonnet’s descriptions of hallucinations, in his original patient, his grandfather Lullin, were described as “…like tapestries” and “… as tableau scenes”, terms that suggest rather flat schematic images. Also reported were fabric patterns, flowers and leaves, a white satin background with black shapes, and circles, all like images on paper. A frequently reported hallucination was of blue handkerchiefs sometimes called squares (“carreaux”, p. 8, 10).
In discussing Bonnet’s syndrome, Morsier (1967) describes one 19th century patient who saw “…posters on the walls of his room” (p. 681); another 20thcentury subject (p. 685) reported flat figures of men and animals. Morsier (1969) also reported patients’ descriptions of hallucinations like “…handkerchiefs,” “…black and white images as in movies,” and other reports strongly suggestive of the schematic, flat, cartoon-like qualities proposed for Stage 3 processing in our model.
Schultz, et al (1996) stated that the face of a person or complex scenery was reported by over 70% of Bonnet’s patients. It is clear from these reports that the images were often quite realistic, and thus can be considered Stage 4 (that is, 3-dimensional) according to our model.
It may be somewhat difficult to accept our claim that individuals can have direct phenomenological access to the neurological structures involved in the internal stages of visual processing. However, the aforementioned evidence certainly implies that this is indeed the case. In normal visual experience, the underlying abstract elements at each stage are assumed to be completely captured by the properties of the natural, external visual stimulus. However, in the absence of any externally initiated visual stimulation, humans appear to have the ability to consciously experience the internally initiated activity of the underlying visual “elements”; this accounts for the hallucinations that are expressed at a particular stage.
The Model: Producing the Image
The model, as described thus far, posits processing of input beginning at a sensory level and progressing to the construction of a phenomenally realistic image at Stage 4. There is, however, another aspect of the model, in addition to input-processing. It seems necessary that in order for a person to draw, there must also be an output-processing mechanism. After all, when a three year old child begins to draw he or she is presumably seeing the world at Stage 4 with all of its 3-dimensional realism. Likewise, many adults with normal 3-dimensional perception are unable to draw or paint at a level as sophisticated as Stage 4. Therefore, although apparently all adults can at least make stick figures, (i.e. early Stage 3 images) we consider that the problem, for children and many adults, is an inadequately developed output or “drawing” module. In particular, they have poor or non-existent Stage 4 output responses.
We propose central nervous system mechanisms that entail output strategies which allow a person to draw a picture. Such proposed output mechanisms are presumably linked to the input processing system, but in many ways also remain distinct. Various theorists have proposed that the deficits in children’s drawings are not caused by conceptual inadequacies – and certainly not perceptual limitations – but instead are the result of production or drawing-skill deficits (see Freeman, 1980; Freeman and Cox, 1985; Cox, 1992, 1993; and Milbrath, 1998).
Perhaps at the highest level of input processing the cortical area devoted to image recognition is roughly similar to the primary speech recognition area (i.e. Wernicke’s). If so, we posit that the image drawing module would be analogous to the speech production area (i.e. Broca’s). We suggest that this output processing model is hierarchical, and that it is organized in the same four stages as the input-processing model. There is research to support a distinction between the neural substrates of visual perception and those underlying visual control (Goodale and Milner, (1992); Goodale (1996). This distinction may involve complex interactions between spatial vision and output responses. I f both input and output hierarchical models exist, then (1) distinct development of these substrates might be evidenced in both perception and drawing patterns during ontogeny; and (2) CNS insult could differentially affect either or both networks to evoke different patterns of perception or drawing either separately or possibly at the same time.
Systematic investigations and theorizing about children’s drawings have been carried out by many. One especially comprehensive study was published by Kellogg (1970), who described progressive development, basing herself upon observation of thousands of drawings made by children from a wide variety of cultural and ethnic backgrounds. She concluded that children’s drawings on average move through the four developmental stages as posited by our model. At first there are onlyscribbles (e.g. Figure 2). Kellogg proposed the next level to involve “basic scribbles.” She identified twenty of these basic scribbles, or as we call them “basic forms”. After the dot, which we place at Stage 1, the first simple forms are straight lines of vertical, horizontal and diagonal orientation, then a curved line, then multiples of these elements. This is followed by a roving open line; then a zigzag, a loop, a spiral and circles. These are shown in Figure 3. Most of these basic forms are typicallyapparent by the time a child is three years old.
At this point, a few comments about the basic forms are in order. Kellogg claimed that dots and next lines of different orientation were the initial basic elements. However, one should keep in mind that not every child develops all the basic forms before moving to the next stage or level. There can be considerable overlap in stages. After basic forms, Kellogg claims that the next elements to develop are “diagrams”, examples of which include the cross shape, square, circle, triangle, X-shape, and a closed figure which is asymmetric and is usually a roving line coming back on itself.
Kellogg strongly emphasized that the level of simple geometric forms develops after scribbles and prior to simple schematic pictorials. Recent scholarship of children’s drawings has been critical of the proposed universality of Kellogg’s more complex geometric shapes, especially mandalas, radials, etc. However, the presence of simple forms similar to Kellogg’s are accepted as a necessary precursor to even early figure type drawings – the first pictorials of children (see Figure 4-a). For example, Gardner, a cognitive psychologist with an extensive knowledge of children’s drawings (Gardner, 1973, 1980), accepts four basic stages of development. He labels the four as scribbles, forms, things (simplified objects), and finally attempts at realism. More recently, Malchiodi (1998) reviewed much of the published work on children’s art and reported the following stages, all of which are compatible with the present model: scribbling, basic forms (simple abstract geometric shapes), human forms and early schemata, followed by more developed schemata and last realism. In children’s art the schematic representation of the human form rapidly becomes more detailed and specific, and soon includes hands, arms, and legs (see Figure 4-b); other elements are sometimes added, for example clothing to indicate sexual differences. This level of representation is often called early pictorialism. At about the same time, the child begins to draw trees, animals and buildings. As noted, all of these theorists propose that all children go through these stages, arriving at a pictorial level at which drawings are flat and schematic. (See also Milbrath, 1998). Only a small number of children, especially the most talented, move on to our Stage 4, where more realistic 3-dimensional drawing involving the use of depth cues commonly makes its appearance.
In Figure 5 this hierarchical process in the development of children’s drawing is summarized by Kellogg in a way that even she considered over simplified, but which nevertheless captures the basic idea clearly.
It is important to note that drawing, at least up to early pictorialism, is not the result of the child looking at something and then attempting to draw it. Instead, the child seems to be attempting to draw some pattern or schema that exists as an internal mental construct. Children’s drawings most of the time do not appear to be based on some object in their external visual environment which they are trying to reproduce. Instead their drawings are determined by a pre-existing schema and by the drawing itself. This similar to how an adult constructs a doodle. Within the framework of our model, evidence for Stage 1 is found in the early scribble drawings and in the use of dots. Stage 2 is shown in basic diagrams and other more complex geometric shapes and designs; and Stage 3 is represented by early and late pictorialism with its flat, cartoon-like and very schematic representation of objects..
There are a few curious exceptions to this approximate developmental sequence. For example, is the well known case of Nadia, an autistic child who demonstrated extremely advanced depictions of depth and realism, almost from her first efforts at drawing at age three to four.( Winner, 1982, pp. 180-188) However, Nadia’s unusual ability was probably closely related to her autism and thus cannot serve as a representative contradiction to the normal progression of stages. For instance, Nadia had eidetic imagery combined with a very limited ability to generalize, form abstractions, and classify objects into categories (Selfe, 1977). Winner noted Nadia “could not match pictures of the same objects unless both pictures represented the object in the same orientation” (p. 186). Hence, we conclude that her case is more of an outlier; the bulk of the evidence supports that four-stage visual and output processing occurs in children.
Alzheimer’s disease is a slow progressive degenerative condition that involves the hippocampus and cortico-temporal networks that are primarily involved in short-term memory consolidation and cognitive processes. Given the protracted nature of this pathology, it is possible to obtain drawings from the same patient over the course of the disease. Within the premises of our model as a degenerative condition affects the networks that subtend general perceptual or perceptual-motor function, the patient’s drawing should evidence “regression” through Stages 4 & 3 to Stages 1 & 2 of the model as higher processing function is successively lost. An example of this is presented in Figure 6, in which the patient was an artist (Cummings & Zarit, 1987). The first drawing is a naturalistic representation and can be described as primarily a Stage 4 image.
The second drawing shows the schematic Stage 3 illustrated by a clear reduction of depth. The last drawing reveals strong geometric Stage 2 properties, although minor aspects of Stage 3 still persist. Judith Wald (1984) presents an obvious case of such stage regression in a female Alzheimer’s patient as shown in Figure 7 (a, b, c). We do not assume that every drawing made after another will be more regressed that the an earlier drawing since the disease can stabilize and can sometimes even appear to improve for short periods of time, but over a period of a year or so regression should be apparent.
A book on art by Alzheimer’s patients (Abraham, 2005) identifies four characteristics of their art work. The first is “simplification” which is a term that could be used in general way to describe regression to our stage 1; however, we describe simplification more specifically. Two other descriptors by Abraham are “fragmentation” and “distortion” which we do not deal with but both of these can be understood as natural consequence in our model of moving from a naturalistic scene through stages 3 & 2 to 1. Her last characteristic is “perseveration” which seems quite reasonable but we have nothing to say about this temporal response.
Another recent representation of regression can be seen in the self portraits of the professional artist William Utermohlen who was first diagnosed with Alzheimer’s disease in 1995. Figure 8a, 1996, shows considerable realism and depth via shading. By 1997 much depth and realism has dropped out and the work shows a primarily schematic, Stage 3 face with a good deal of Stage 2 geometric patterning in the background; in 1999 the face is now primarily Stage 1. He died of the disease in 2007.
Some final support comes from the late paintings of the prominent abstract expressionist artist Willelm De Kooning. (For information on De Kooning’s life and late works see Storr and Garrels, 1997. Examples of his late paintings made in the late 1980’s can be easily found by going on-line under his name and “late paintings” In De Kooning’s earlier and most representative period his works typically showed recognizable parts of a woman’s body, especially aspects of the face with the rest of the work being without pictorial content. However, a few of his works during this mature period were also without any imagery or even geometric emphasis. He was diagnosed with probable Alzheimer’s disease around 1980 and he also suffered from the effects of alcoholism based on many earlier years of heavy drinking. He continued to paint until mid 1990 and eventually died in 1997. From the perspective of the present model all of De Kooning’s works of the late period are without any Stage 3 schematic imagery and without any systematic geometric patterning, Stage 2. All are either wandering line like works usually with strong red, blue and yellow or orange and occasional green color in the line or in shapes created by such relatively thick wandering lines. A few others are just dark areas of color without any line properties. In short, all of his quite late paintings appear to be at Stage 1. In a modest way some of his late works were prefigured by aspects of some of his pre-Alzheimer paintings but his late works are, in general, quite distinctive. This interpretation is not to deny that they have artistic merit. De Kooning was an extremely talented and innovative artist but he seems to have been driven down to this lowest level, according to the model, by his Alzheimer’s condition. Nevertheless, even here he was able to express an intriguing aesthetic. Louis Wain, a much smaller artist who suffered from severe mental pathology, discussed below, produced his most artistically interesting works when he had regressed from Stage 4 to earlier stages. Again as mentioned earlier the stage level of a work is not a comment on its aesthetic value.
We also expect that a similar representational regression might be found in some stroke patients or other brain damaged artists such that the greater the (network hierarchical) effect of the CVA, the greater the regression of the patient’s ability at visual representation. We posit that progressive recovery from the damage would coincide with a reversal of such change(s), and advancement to higher stages of processing and representative output. One example supporting
this proposal is found in the progressive recovery of a brain damaged artist provided by Zaidel (2005 p. 26-28) in Figs 2.1a thru 2.1f. The artist’s portraits move from very schematic, even child-like in 2.1a to flat and cartoon-like 2.1b to real depth 2.1g.
Psychiatric pathology and recovery
Before we address psychiatric pathology and recovery, it is important to summarize the earlier research on psychiatric hallucinations conducted by Horowitz (1964). In this study, subjects reported examples of spots, grids, wavy lines, circles, radiating figures, and similar Stage 2 types of hallucinations; although “amorphous specks” (Stage 1), schematic shapes of a faucet, and animal heads (Stage 3), realistic animate objects such as a parent or sibling (Stage 4) were also described. Horowitz contends that when interpreting the reported hallucinations, (especially those from anxious and depressed psychiatric patients), “…it was necessary to persist beyond initial verbal descriptions … and insist that the patient describe and draw what he had seen. Initial descriptions of vicious snakes might then be drawn and re-described as wavy lines. In one case the patient’s description of ‘Two armies struggling over my soul’ arose from the subjective experience of seeing moving sets of dots.” (p. 513). In short, psychiatric patients (and probably non-psychiatrically disordered individuals) often provide fairly complex narrative interpretations of what are, in fact, much simpler visual percepts.
Serious psychiatric conditions can also result in a loss of mental capacity that damages adults’ ability to draw. Thus, we can expect that the severity of psychiatric disturbance will show up in changes to the stage or levels of a person’s drawing. Prior to any evidence of mental pathology, a typical adult should be drawing at either Stage 3 or 4 depending upon his or her skill. As a mental condition becomes more severe, regression through the stages should become evident (as analogous to the effects of Alzheimer’s disease). A number of examples support this possibility; perhaps best known is the work of artist Louis Wain. During his normal adulthood he was well known for his paintings and drawings of cats, which appeared in many illustrated English periodicals (Dale, 1991). These were realistic and the cats were often shown in humorous situations, implicitly mimicking human interactions. After some years as a successful artist however, Wain became mentally disturbed and was hospitalized for schizophrenia. The rather well known series of five drawings that demonstrate the regression in Wain’s work can be reliably found under his name on the internet. (E.g. Cardoso, no date) In the series of his cats presented by Cardoso the last example is devoid of any recognizable content and is essentially only a Stage 2 image from the perspective of the present model. (Cardoso’s use of the word “stage” is different from ours.) It is clear that the early signs of deterioration are present by the changes noted between the first and third drawings: the third ‘cat’ shows considerable loss of depth (via shading) and there is the emergence of much schematic, cartoon-like imagery. The loss of depth is the major cue to movement from Stage 4 down to Stage 3.
There is no reliable evidence that the usual series of Wain’s cats were actually made in the order shown but that seems somewhat irrelevant. Over the years schizophrenia, like most mental pathologies, can wax and wane for a given patient, even if the general tendency is one of decline over time. In any case, the deterioration in the images is obvious and Wain’s changes are supported by the other examples shown here where the mental state of the artist at the time of each drawing is known. It is interesting to note that when Wain’s images have been shown to students in various university classes by the author (P. C. V.) it was often remarked by art students that the aesthetic value of Wain’s work became greater and the drawings more interesting in the early stages of his regression. Another clear example of such representational regression in a mental patient can be found in E. C. Dax, 1998, p.83 where the artist/patient is listed under schizophrenia; we are informed he suicided not long after the last more stage regressed drawing.
Recovery from mental disturbance, however, should result in, and reveal, reconstitution of the four-stage hierarchy of representational capacity. An example of this is shown in the works of “Mr. Pauli” (Figure 9). As reported by R. M. Simon (1997), Pauli began painting in a psychiatric group; he subsequently left the group, was discharged from the hospital, and lived on his own. He apparently became progressively “more normal”, as visible in the sequence of his works (which were left at various times at Simon’s door). The last picture Pauli brought to her is number 9-d, a robust representation of a Stage 4 drawing.
A student of the first author who was working in a hospital observed a patient who showed stage recovery in her drawings as shown by Figure 10. The drawings were done while the patient was in a short-term psychiatric ward in New York City. The first day she was admitted she drew 10-a; the next day 10-b and the third day, she drew 10-c, shortly before being discharged.
Figure 11 presents the drawings of another patient admitted to the same short-term psychiatric ward. Drawings 11 a, b, c show systematic stage regression and were produced on the first, third, and fourth day after his admission respectively; a day later he was transferred to a long-term psychiatric ward. These observations suggest that stage level evaluations of a person’s drawings may be useful for getting a quick measure of the degree of pathology.
Visual System Location
There is a modest but relevant literature on the anatomical localization of processing of phosphenes and related visual images in humans. Chapanis, Uematsu, Konigsmark and Walker (1973), and Brindley (1968) report the sensations induced by stimulation of the visual cortex of a recently blinded woman. It was shown that stimulation using a simple cortical electrode usually led to the sensation of a single spot of light at a constant location in the visual field. It was also reported that simultaneous stimulation by several electrodes caused the patient to perceive predictable simple patterns. (These findings support our model, and proposed mechanisms of Stages 1 and 2). Chapanis, et al (1973) studied three patients implanted with thalamic electrodes via the occipital lobe. Electrical stimulation of the electrodes led to the perception of phosphenes described as round, square, triangular or rectangular. (This is evidence for Stage 2). Lance (1976) reports that Foerster (1931) claimed that unformed visual hallucinations, such as flashes of light, originated in the retina or primary receptive area for vision in the occipital cortex (Brodmann’s areas 17 and 18); and also reported that organized images were obtained by stimulation of area 19. Lance summarized reports of his patients who had parieto-occipital lesions and concluded that their perceptions consisted primarily of objects, people, and animals, and that those were presumably “…the common building blocks of visual memory, a matrix for the channeling of visual memory rather than the visual memory itself.” These images were more complex than the “…flashes, zigzags and wholes of light and color obtained from the primary visual cortex…” (all quotes p. 732). These reports provide evidence for Stages 1, 2 and 3 of visual processing, as consistent with hierarchical mechanisms progressing from lower to higher levels. Finally, a review by Ffytche and Howard (1999) summarized the hallucinations of 50 patients with degenerative ocular disease, and reported that the patients’ hallucinations were rather stereotypical, and that the visions were likely due to de-stimulatory, indirect neuropathological changes of the visual cortex rather than pathology of the eye, directly. The ocular diseases were senile macular degeneration (58%), glaucoma (18%) and a variety of other pathologies (24%). Given the different pathological etiologies of blindness and the absence of any focal cerebral pathology, one should not expect a simple connection between the type of hallucination and the CNS-dependent stages described by our model. Yet all of the reported images are representative of one of the four Stages of the model – although the effects of movement, size change, etc., appear to be unrelated. Reports of visual phenomena included “… fireworks exploding in vivid color” (Stage 1), “…nets in sharp geometric shapes”, other patterns (Stage 2), multiple copies of objects (e.g. rows of mugs, faces distorted or cut-up like an early Picasso ( Stage 3) and a number of naturalistic images (Stage 4).
Is this merely serendipitous, or do these perceptions support distinct CNS effects of differing visual activity? To propose an answer, we address where the model’s stages might engage local network activities in the brain. Stage 1, we propose takes place in the retina, the lateral geniculate nucleus (LGN), the primary visual cortex (A-17), and is probably completed at the color center found at V4 (Zeki, 1990) Thus, this stage is presumed to reflect relatively low level neuro-cortical events. It is well known that color, brightness, and spot- or point- shaped receptive fields are found in the retina and LGN and that V4 is a kind of final stage color center. We posit that these structures constitute the primary neurological substrates for Stage 1 processing. Studies by Grüsser, Grüsser–Cornehls, Kusel and Przybyszewski (1989), seem to support this contention.
We believe that Stage 2 mostly engages the primary or initial areas of the visual cortex, where orientation, straight lines and angles are detected, e.g. V1 – V3, and probably somewhat higher visual areas and possibly a final center, analogous to the Zeki color center, which organizes abstract geometric shapes.
Stage 3 processing is assumed to take place and be completed in as yet unidentified higher levels of the cortex, although various studies implicate the lateral-occipital complex (see below). More specifically, Stage 3 can be interpreted as engaging an object recognition area. This means that objects are first recognized in schematic, simplified form, and that this takes place after Stage 2 processing, and before Stage 4.
Research on object recognition supports the idea that the proposed Stage 3 engages higher, and perhaps multiple cortical and extra-cortical- networks. (See, for example, Malach, et al, 1995.) In general, there is reliable evidence that there are two visual processing systems, one for recognizing forms/object (“what” is seen) and the other for locating the object in the visual space (“where” it is seen) (Ungerleider, Courtney, & Haxby, 1998). Object recognition is processed in the ventral visual network – probably involving the lateral-occipital complex (Kanwisher, Woods, Iacoboni, & Mazziotta, 1997; Haxby, Ungerleider, Clark, Schouten, Hoffman, & Martin, 1999). Doniger and colleagues (2001), used schematic representations (e.g. simple cartoon-like representations of an elephant), and noted that this type and level of recognition takes place in the lateral-occipital complex, secondary to early cortical visual processing (e.g. V1 – V3) where we presume most of Stage 2 occurs.
It is important to note that in this model, the actual location for object recognition need not be specified, but only that the process involves hierarchical networked activations. There is an extensive body of research on object recognition that has used schematic or cartoon-like drawings of objects which implicitly supports the presence of Stage 3. For example, the line drawings of Snodgrass and Vanderwart (1980) support this hypothesis. There is also evidence that the human visual system can recognize objects from partial information (Snodgrass and Feenan, 1990). This capacity, called “perceptual closure”, implies that simple object schemas are present in the visual system, and can be activated by incomplete (object) information. We hold that this strengthens our concept of Stage 3 processing.
Our model of Stage 3 and (possibly) Stage 4 processing, as noted, receives further support from studies that have shown object recognition and spatial vision to be subtended by separate cortical loci/networks. Presumably 3-D spatial representation and localization of an object occurs at Stage 4, after earlier schematic or simplified object recognition. (See Kohler, Kapur, Moscovitch, Winocur, Houle,1995; Moscovitch, Kapur, Kohler, Houle, 1995 for discussion.) The present model, in its simplest form, implies that the perceptual “construction” at Stage 3 does not include the functional or associative meaning of an object. Presumably these properties are processed at higher levels of the networked hierarchy that are different from, and probably subsequent to, the perceptual construction of the object per se.
Our input-processing model predicts that mental pathology, (e.g. schizophrenia, or neurological deterioration, e.g. Alzheimer’s), could insult and affect the input processing at one or more stages – not just the systems that are involved in output or drawing. That is, if the drawing ability of a patient suffering from mental pathology, regresses to Stage 3 or 2, then this deterioration in output might be accompanied by and reflect deterioration of the related perceptual processing. Support for this interpretation is provided by Doniger, et al, (2001) who report that schizophrenics have object recognition deficits that are likely due to impaired sensory/perceptual processing.
Some four decades ago, Klüver (1966, as cited by Ffytche and Howard) observed commonalities in abnormal visual experience. Such constants suggested “… some fundamental mechanisms involving various levels of the nervous system” (Ffytche & Howard, 1999, p. 1255). The present model, while focusing upon artistic perception and output (i.e. drawing), reinforces and extends these findings.
Parallel and Serial Processing
It is well established that parallel processing underlies much visual perception. Certainly, within a stage (of our model), parallel processing can, and we presume, commonly does occur. For example, if we consider Stage 1 to be primarily subtended at retinal and LGN levels, then the parallel processing of these substrates- as contributory to Stage 1 perception- is well known. The actual relationships of neural networks involved in higher levels of the model, as far as we are aware, remain unknown, although the connection of Stages 1 and 2 can be considered a function of the visual radiations engaging (the highest) Level 1 processing onto Stage 2.
It is important to note, however, that serial processing also operates within this model: the four qualitative stages are explicitly predicted to occur in a hierarchical sequence. That is, the gross structure of the model involves hierarchically qualitative stages that are, at the macro level, serial. The qualitative stages in the model are consistent with the known fact that cortical areas have an identifiable hierarchy based on laminar patterns of inputs and outputs. There is strong evidence for certain types of serial processing in the cortex. (See Doeringer & Hogan,1998; Hubel & Wiesel, 1965, 1977; Inui, Wang, Tamura, Kaneoke & Kakigi,2004: Pons, Garraghty, Friedman & Mishkin, 1987.) Nevertheless, as noted, especially within a given stage, the model is assumed to predominantly involve parallel processing. Still, it is reasonable to presume that parallel and serial processing occur synergistically both within and between levels of hierarchy. This is consistent with hierarchical cognitive function as proposed by Jackendoff (1987) and extended by Prinz (2000).
Bottom Up and Top Down Processing Networks
As presented, the model could be seen as a “bottom up” processing model. But, we accept the recently proposed interpretations that cortical processing commonly involves bi-directional networked hierarchies, such that prominent bottom-up and top-down effects are reciprocal (Bar, 2004, Hawkins, 2004, Prinz, 2000). The present model allows this bi-directionality within and between stages. For example, at Stage 2 ascending and descending networks are presumed to be involved in the perception/construction (and production) of lines and geometrical patterns initiated by the stimulus; the descending projections may involve maintenance of the accuracy of stimulus input so as to strengthen, focus, and perhaps contextualize it to other dimensions of cognition and/or emotion. This is consistent with the cited models and would reflect complex dynamical systems activation and integration in the CNS.
The Model and Primitive Art
Animals do not draw, unless artificially encouraged to do so by humans – and even then they show no reliable evidence for drawing recognizable objects. Therefore we, like others, assume that making designs and drawings constitute distinctly human activity. It is also likely that drawing is closely connected to human language. Indeed the existence of designs, much less drawings, have been commonly used as evidence for “identifying” the first humans, that is, the first primates capable of human consciousness and capable of symbolic thought and expression. There is much debate over when and where this occurred. Some have argued for a relatively recent date of around 32,000 BCE (Pfeifer, 1982), and others for around 75,000 BCE (Hogan, 2003). A few claim that very simple marks are examples of even earlier human art. The very earliest claims are for “cupules” (Bednarik, 1993), which are small hemispherical indentations pounded into flat, sloping or vertical rock surfaces. Cupules are very analogous to the present model’s understanding of the simplest and earliest element, the dot. However, when cupules are found without other more complex markings the claim that they represent early art seems to stretch the concept of art to a breaking point. Regardless, for present purposes, it is a common observation that in the case of primitive art, very simple elements appear first, then geometric patterns, then more complex designs and finally stick drawings and schematic objects. More complex realism comes still later, if at all. (See also Zaidel,2005)
In concluding we summarize. The present model is a four-stage-hierarchical description both of how an image is processed or constructed to visual completion and also of how it is produced when the image is drawn. Supporting evidence is drawn from the following: over a century of published reports of visual hallucinations; the extensive literature on the development of children’s drawings; and artistic regression in persons suffering from mental or medical pathology. In addition, much recent research in cognitive neuroscience has been cited to connect the model, at least in a general way, to brain functions. The model may also bear on historical development or changes in art such as interpretations of the origin of art in early human societies.
Finally, this paper by focusing on the relevance of a patient’s art interprets the patient as a whole person, not just a collection of parts, thus emphasizing medicine as an integration of both the sciences and the humanities.
Abraham, R. (2005). When words have lost their meaning: Alzheimer’s patients communicate through art. Westport, CT: Praeger.
Bar, M. (2003). A cortical mechanism for triggering top-down facilitation in visual object recognition. Journal of Cognitive Neuroscience, 15, 600-609.
Bexton, W. H., Heron, W., & Scott, T. (1954). Effects of decreased variation in the sensory environment. Canadian Journal of Psychology, 8 (2), 70 – 76.
Brindley, G. S., & Lewin, W. S. (1968). The sensations produced by electrical stimulation of the visual cortex. Journal of Physiology, London, 196, 479 – 493.
Chapanis, N. P., Uematsu, S., Konigsmark, B., & Walker, A. E. (1973). Central phosphenes in man: a report of three cases. Neuropsychologia, 11, 1-19.
Cole, M. G. (1992). Charles Bonnet hallucinations: a case series. Canadian Journal of Psychiatry, 3, 267 – 270.
Cardoso, S. H. (no date). Cats painted in the progression of psychosis of a schizophrenic artist. Neuroscience Art Gallery, http:www.cerebromente.org.br/gallery/gall-leonardo/fig1-a.htm Nov. 3, 2011.
Cox, M. (1992). Children’s Drawings. London: Penguin.
Cox, M. (1993). Children’s Drawings of the Human Figure. England: Erlbaum.
Cummings, J. L., & Zarit, J. M. (1987). Probable Alzheimer’s disease in an artist. Journal of American Medical Association (JAMA), 258 (19), 2731 – 2734.
Dale, R. (1991). Louis Wain: The man who drew cats. London: Michael O’Mara Books.
Dax, E. C. (1998). Schizophrenia. In The Cunningham Dax Collection: Selected works of psychiatric art (pp. 55 – 88). Carlton South, Victoria: Melbourne University Press.
Doniger, G. M., Foxe, J. J., Schroeder, C. E., Murray, M. M., Higgins, B. A., & Javitt, D. C. (2001). Visual perceptual learning in human object recognition areas: A repetition priming study using high-density electrical mapping. NeuroImage, 13,305 – 313.
Doniger, G., Foxe, J., Murray, M., Higgins, B., Snodgrass, J. G., Schroeder, C. E., & Javitt, D. C. (2000). Activation time course of ventral visual stream object-recognition areas: High density electrical mapping of perceptual closure processes. Journal of Cognitive Neuroscience, 12 (4), 615 – 621.
Doniger, G., Silipo, G., Rabinowitz, E., Snodgrass, J., & Javitt, D. (2001). Impaired sensory processing as a basis for object recognition deficits in schizophrenia. American Journal of Psychiatry, 158, 1818-1826.
Doeringer, J. A. & Hogan, R. (1998). Serial processing in human movement production. Neural Networks, 11, 1345-1356.
Flournoy, T. (1902). Le case de Charles Bonnet. Hallucinations visuelles chez un vieillard apere de la cataracte, Archives de Psychologie (Geneve), 1 (1), 1 – 23.
Foerster, O. (1931). The cerebral cortex in man, Lancet, ii, 309 – 312.
Freeman, N. H. (1980). Strategies of representation in young children. London: Academic
Freeman, N. H., & Cox, M. (1985). Visual order. London: Academic Press.
Ffytche, D. H., & Howard, R. J. (1999). The perceptual consequences of visual loss: ‘positive’ pathologies of vision, Brain, 122, 1247 – 1260.
Gardner, H. (1973). The arts and human development. New York: Basic Books.
Gardner, H. (1980). Artful scribbles. New York: Basic Books.
Golomb, C. (2002). Child Art in Context: A cultural and comparative perspective. Washington, DC: American Psychological Association.
Goodale, M. A. (1996). Visuomotor modules in the vertebrate brain. Canadian Journal of Physiological Pharmacology, 74 (4), 390 – 400.
Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neuroscience, 15 (1), 20 – 5.
Grüsser, O. J., Grüsser-Cornehls, U., Kusel R., & Przybyszewski, A. W. (1989). Responses of retinal ganglion cells to eyeball deformation: a neurophysiological basis for ‘pressure phosphenes.’ Vision Research, 29 (2), 181 – 194.
Grüsser, O. J., & Hagner, M. (1990). On the history of deformation phosphenes and the idea of internal light generated in the eye for the purpose of vision. Documenta Ophtalmologica, 74, 57 – 85.
Hawkins, J. with Blakeslee, S. (2004). On intelligence. New York: Times Books/Henry Holt.
Haxby, J. V., Ungerleider, L. G., Clark, V. P., Schouten, J. L., Hoffman, E. A., & Martin, A. (1999). The effect of face inversion on activity in human neural systems for face and object perception. Neuron, 22, 189 – 199.
Heron, W. (1957). The pathology of boredom. Scientific American, 196,52 – 56.
Heron, W. (1961). Cognitive and physiological effects of perceptual deprivation. In Solomon et al (Eds.), Sensory Deprivation (pp. 6 – 33). Cambridge: Harvard University Press.
Höfer, O. (1963). Über die Abhänigkeit elektrisch induzierter, subjectiver Muster (Phosphene). von zeitlichen Verlauf der anregenden Wechselströme. Elektomedizin, 8 (2), 72 – 87.
Hogan, J. (2003, April, 18). Doubt cast on age of oldest human art. New Scientist, #2391.
Horowitz, M. J. (1964). The imagery of visual hallucinations. The Journal of Nervous and Mental Disease, 138(6), 513 – 523.
Hubel, D. H. & Wiesel, T. N. (1965). Receptive fields and functional architecture in two non- stiate visual areas (18 and 19) of the cat. Journal of Neurophysiology, 28, 229-289.
Hubel, D. H. & Wiesel, T. N. (1977) Functional architecture of macaque monkey cortex. Proceedings of the Royal Society of London, 198, 1-59.
Hufford, J. (1983). An overview of developmental stages in children’s drawings. In M. Zurmuehlen (Ed.), Working Papers in Art Education, n 2 (pp. 10-15). (Available from Working Papers in Art Education, 13 North Hall, The University of Iowa, Iowa City, IA 52242).
Inui, K., Wang, X., Tamura, Y., Kaneoke, Y. & Kakigi, R. (2004). Serial processing in the human somatosensory system. Cerebral Cortex, 14, 851-857.
Jackendoff, J. (1987). Consciousness and the computational mind. Cambridge, MA: MIT Press.
Kanwisher, N., Woods, R. P., Iacoboni, M., & Mazziotta, J. C. (1997). A locus in human extrastriate cortex for visual shape analysis. Journal of Cognitive Neuroscience, 9,133-142.
Kawabata, H. & Zeki, S. (2004). Neural correlates of beauty. Journal of Neurophysiology, 91, 1699-1705.
Kellogg, R. (1970). Analyzing Children’s Art. Palo Alto, CA: National Press Books.
Kellogg, R., Knoll, M., & Kugler, J. (1965). Form-similarity between phosphenes of adults and pre-school children’s scribblings. Nature, 208 (5015), 1129 – 1130.
Klüver, H. (1966). Mescal and mechanisms of hallucinations. Chicago: University of Chicago
Knoll, M., Kugler, J., Eichmeier, J., & Höfer, O. (1962). Note on the spectroscopy of subjective light patterns. Journal of Analytical Psychology, 7 (1), 55 – 69.
Knoll, M., Höfer, O., & Kugler, J. (1966). Fliegerpersönlichkeit und Phosphenhäufigkeit. Beiheft zur Schweiz: Zeitschrift fur Psychologie und Ihre Anwendungen, 50, 253 – 281.
Knoll, M., & Welpe, E. (1968). Vergleich von Anregungsbedingungen, For-Klassen und Bewegungsgarten optischer und elektrischer Phosphene. Elektromedizin, 13 (4), 128 – 134.
Kohler, S., Kapur, S., Moscovitch, M., Winocur, G., & Houle, S. (1995). Dissociation of pathways for object and spatial vision: A PET study in humans. Cognitive Neuroscience and Neurophysiology, 6, 1605 – 1615.
Lance, J. W. (1976). Simple formed hallucinations confined to the area of a specific visual field defect. Brain, 99, 719 – 734.
Malach, R., Reppas, J. B., Benson, R. R., Kwong, K. K., Jiang, H., Kennedy, W. A., …& Tootell.(1995). Object-related activity revealed by functional magnetic resonance imaging in human occipital cortex. Proceedings of the National Academy of Sciences, U.S.A., 92, 8135 – 8139.
Malchiodi, C. A. (1998). Understanding children’s drawings. New York: Guilford.
Milbrath, C. (1998). Patterns of artistic development in children: Comparative studies of talent. New York: Cambridge University Press.
de Morsier, G. (1967). Le syndrome de Charles Bonnet: hallucinations visuelles des vieillards sans deficience mentale. Annales Medico-Psychologiques, 12 (5), 677 – 702.
de Morsier, G. (1969). Les hallucinations visuelles diencephaliques: deuxieme partie. Psychiatria Clinica, part 1: 167-184; part 2: 232 – 251. Basel: S. Karger.
Mortensen, K. V. (1984). Children’s Human Figure Drawings (Vol. 2). Copenhagen: Dansk Psykologisk Forlag.
Moscovitch, C., Kapur, S., Kohler, S., & Houle, S. (1995). Distinct neural correlates of visual long-term memory for spatial location and object identity: A positron emission tomographynstudy in humans. Proceedings of the National Academy of Sciences, U.S.A., 92, 3721 – 3725.
Oster, G. (1970). Phosphenes. Scientific American, 222 (2), 82 – 87.
Pfeifer. J. E. (1982). The creative explosion: An inquiry into the origins of art and religion. New York: Harper & Row.
Pons, T. P., Garraghty, P. E., Friedman, D. P. & Mishkin, M. (1987). Physiological evidence for serial processing in somatosensory cortex. Science, 237, 4813, 417-420.
Prinz, J. J. (2000). A neurofunctional theory of visual consciousness. Consciousness and Cognition, 9, 243-259.
Schultz, D. P. (1965). Sensory Restriction: Effects of behavior. New York: Academic Press.
Schultz, G., & Melzack, R. (1991). The Charles Bonnet syndrome: ‘phantom visual images.’ Perception, 20,809 – 825.
Schultz, G., Needham, W., Taylor, R., Shindell, S., Melzack, R. (1996). Properties of complex hallucinations associated with deficits in vision. Perception, 25,715 – 726.
Seidel, D., Knoll, M., & Eichmeier, J. (1968). Anregung von subjektiven Lichterscheinungen (Phosphenen) beim Menschen durch magnetische Sinusfelder. Pflügers Archives, 299, 11 – 18.
Selfe, L. (1977). Nadia: A case of extraordinary drawing ability in an autistic child. London: Academic Press
Simon, R. M. (1997). Mr. Pauli, madly serious play. Chapter 2 in Symbolic Images in Art as Therapy (pp. 15 – 37). London: Routledge.
Snodgrass, J. G., Feenan, K. (1990). Priming effects in picture fragment completion: Support for the perceptual closure hypothesis, Journal of Experimental Psychology: General, 119, 276 – 296.
Snodgrass, J. G., Vanderwart, M. (1980). A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity, Journal of Experimental Psychology: Human Learning and Memory, 6, 174 – 215.
Storr, R. and Garrels, G. (1997). Willem deKooning: The late paintings: The 1980s. San Francisco, CA: San Francisco Museum of Modern Art.
Tyler, C. W. (1978). Some new entropic phenomena, Vision Research, 18, 1633 – 1639.
Ungerleider, L. G., Courtney, S. M., & Haxby, J. V. (1998). A neural system for human visual working memory, Proceedings of the National Academy of Sciences, U.S.A., 95, 883 – 890.
Vernon, J. A. & McGill, T. E. (1962). Sensory deprivations and hallucinations. In L. J. West (Ed.), Hallucinations (pp. 146 – 152). New York: Grune and Stratton.
Vernon, J. A. (1963). Inside the black room. New York: C.N. Potter.
Wald, J. (1984). The graphic representation of regression in an Alzheimer’s disease patient, The Journal of Art Psychotherapy, 11, 165 – 175.
Wald, J. (1989). Art therapy for patients with Alzheimer’s disease and related disorders. Chapter 11 in H. Wadeson, J. Durkin, & D. Perach (Eds.), Advances in Art Therapy (pp. 204 – 221). New York: Wiley.
Winner, E. (1982). Invented worlds: The psychology of the arts. Cambridge, MA: Harvard University Press.
Zaidel, D. W. (2005). Neuropsychology of art: Neurological, cognitive and evolutionary perspectives. East Sussex, UK: Psychology Press.
Zeki, S. (1990). A century of cerebral achromatopsia. Brain, 113, 1721-1777.
Zeki, S. (1998). Parallel processing, asynchronous perception and a distributed system of consciousness in vision. The Neuroscientist, 4, 365-372.
Zubek, J. P. (Ed.). (1969). Sensory Deprivation: Fifteen years of research. New York: Meredith.
Paul C. Vitz, Department of Psychology, New York University, Professor Emeritus and Professor, The Institute for the Psychological Sciences; Tatiana Kamorina, Department of Psychology, New York University.
We would like to acknowledge colleagues whose critical readings of earlier drafts of this ms. were extremely helpful: At IPS, James Giordano and at NYU, Edgar E. Coons, Denis Pelli and Joan G. Snodgrass, and at NIMH, Leslie Ungerleider.
Correspondence concerning this article should be addressed to Paul C. Vitz, IPS, Suite 511, 2001 Jeff Davis Highway Arlington, VA 22202. E-mail:firstname.lastname@example.org