Balint's Syndrome or placed on a table in front of him he could rarely Coexisting visual field deficits,hemispatial bring his spoon to it at once,but had to grope for it neglect,apperceptive or associative agnosia,pro- till he had located it by touch"(Holmes Horax. sopagnosia,alexia,and other cognitive deficits are 1919.pp.391and393). often present in association with Balint's syndrome or some of its constituent elements Impaired Depth Perception The patient reported by Balint (1909)for example.also had left hemispatial neglect,possibly Holmes and Horax (1919)also attributed impaired owing to extension of the lesion into the right ten depth perception to spatial disorientation. They arietal junction (figure 21):"TJhe attent viewed the loss of depth ception in balint's syr of the patient is always cted [y drome as of the 40 deg esl to the ght-hand side of ption and as a failure to hav loss of topogr he is sked to direct his autention In the t the the los after ha o fixed his of blinking in se to a visual threat to the tends to the right-hand rather left hand ient's inability to re s of the side (cited by and St 1988 90 ing obje ing dis ance objec . llide with objects when they w cy of the patie mpairme in Ba te to ems to ppre may e patien a(Ridd Humphreys, t he ing a the distanc ataxi spatial isorient elements of Bilin syn lack of a se of di e di dis ndicate a la s in genera rk.C ince he could:indicate by hi two hands the exten simulta sion of ordinary standards of linear measuremen or spatial disorientation(Perenin vighetto as an inch.a foot,or a yard. and he could inc and simultanagnosia may occur without spatial dis cate the lengths of familiar objects,as his rifle 1963 bayonet,etc.(Holmes Horax.1919.p.393) Rizzo Robin,1990).It should be bome in mind that in such cases.the observed symptoms may result from very different mechanisms than the Nosological Consideration:Balint's Syndrome that produce them in Balint's syndrome.Thus.while Its Neighbors and Relatives optic ataxia and oculomotor impairment may be attributable to a loss of spatial representation in The clinical picture described here is that of Balint' patients with Balint's syndrome caused by bilateral syndrome when it dens and in s pure parieto-occipital lesions,optic ataxia from superior P parietal lesions may reflect disruption of the neural with bilate e pariet substrates mediating visuomotor transformations ipital junction. Whi stroke and head traum (Milner Goodale.1995). may occasionally cause discretely restrict ed and Similarly.simultanagnosia may be caused by symmetrical le ions,it is more commonly the cas very different kinds of lesions for different reasons that lesions will not respect these territories and The term simultanagnosia was originated specifi- will cause more extensive damage to the occipital. cally to describe a defect in integrating complex parietal,and temporal lobes
or placed on a table in front of him he could rarely bring his spoon to it at once, but had to grope for it till he had located it by touch” (Holmes & Horax, 1919, pp. 391 and 393). Impaired Depth Perception Holmes and Horax (1919) also attributed impaired depth perception to spatial disorientation. They viewed the loss of depth perception in Bálint’s syndrome as a consequence of the loss of topographic perception, and as a failure to have any appreciation of distance. In their patient they attributed the loss of blinking in response to a visual threat to the patient’s inability to recognize the nearness of the threatening object. Difficulty in judging distances also causes another serious problem for patients— they collide with objects when they walk about. The impairment of depth perception in Bálint’s syndrome seems to be due to a failure to appreciate the relative location of two objects, or of the patient and the object he or she is looking at. Size cues seem not to help the patient judge the distance to an object. However, Holmes and Horax commented that their patient’s lack of a sense of distance did not indicate a lack of appreciation of metrics in general since he could: “indicate by his two hands the extension of ordinary standards of linear measurement, as an inch, a foot, or a yard... and he could indicate the lengths of familiar objects, as his rifle, bayonet, etc. (Holmes & Horax, 1919, p. 393). Nosological Consideration: Bálint’s Syndrome, Its Neighbors and Relatives The clinical picture described here is that of Bálint’s syndrome when it is quite dense and in its pure form. It reflects the typical presentation of a patient with bilateral lesions restricted to the parietooccipital junction. While strokes and head trauma may occasionally cause discretely restricted and symmetrical lesions, it is more commonly the case that lesions will not respect these territories and will cause more extensive damage to the occipital, parietal, and temporal lobes. Coexisting visual field deficits, hemispatial neglect, apperceptive or associative agnosia, prosopagnosia, alexia, and other cognitive deficits are often present in association with Bálint’s syndrome or some of its constituent elements. The patient reported by Bálint (1909), for example, also had left hemispatial neglect, possibly owing to extension of the lesion into the right temporoparietal junction (figure 2.1): “[T]he attention of the patient is always directed [by approximately 35 or 40 degrees] to the right-hand side of space when he is asked to direct his attention to another object after having fixed his gaze on a first one, he tends to the right-hand rather than the left-hand side” (cited by Husain and Stein, 1988, p. 90). In other cases in which a constriction of visual attention is also associated with object agnosia, the tendency of the patient to become locked on parts of objects may contribute to observed agnosic errors and may result in diagnostic confusion with integrative agnosia (Riddoch & Humphreys, 1987). It is also the case that a given patient may have optic ataxia, spatial disorientation, or simultanagnosia without other elements of Bálint’s syndrome. Thus, spatial disorientation may occur without simultanagnosia (Stark, Coslett, & Saffran, 1996); optic ataxia may occur without simultanagnosia or spatial disorientation (Perenin & Vighetto, 1988); and simultanagnosia may occur without spatial disorientation (Kinsbourne & Warrington, 1962, 1963; Rizzo & Robin, 1990). It should be borne in mind that in such cases, the observed symptoms may result from very different mechanisms than those that produce them in Bálint’s syndrome. Thus, while optic ataxia and oculomotor impairment may be attributable to a loss of spatial representation in patients with Bálint’s syndrome caused by bilateral parieto-occipital lesions, optic ataxia from superior parietal lesions may reflect disruption of the neural substrates mediating visuomotor transformations (Milner & Goodale, 1995). Similarly, simultanagnosia may be caused by very different kinds of lesions for different reasons. The term simultanagnosia was originated specifi- cally to describe a defect in integrating complex Balint’s Syndrome 33
Robert Rafal visual scenes (Wolpert,1924).As defined by Wolpert,the term includes.but is m Object-and Space-Based Attention neral thar onstriction of attentio in Balint's nd ppr A eciation of simultanag osia in Balint's drome.It is r tha has tial in helpin ut from lateral lesion rch.The ties of their whether visual attentio light for "[W]hen the objec Work away Ir the cig (P 1980P by s was unal ble to se th eyes were 1956R3574H 0n field enh sign t th may be of attention. tulate tha preatter ve pro sses par 1 a in Balint's syndrome andi ate objects (m ore on thi Unlike e in Balint's syndrome,simultanagnosi on then acts b caused by unilateral left temporoparietal lesions r processing that can gud appears to be due to a perceptual bottleneck caused goal-directed action.These models are supported by slowing of visual processing as measured by by experiments in normal individua that shov (e10 better discrmination of two features belonging to the same object than of features belonging to two trast.patients with Balint's syndrome may be able different objects (Duncan,1984)and that thes to recognize a series of individual pictures flashed object-based effects are independent of the spatia briefly in an RSVP test (Coslett Saffran.1991). location of their features (Bavlis Driver.1995: Vecera Farah,1994). Physiological recordings have shown that an plications of Balint's Syndrome for object-based attentional set can modulate process- Understanding Visual Cognition ing in the extrastriate visual cortex (Chelazzi, Duncan.Miller,Desimone,1998).Recent neu- lint's syndrome holds valuable lessons for under roimaging studies have confirmed that attentional tanding the neural processes involved in contro selection of one of two objects results in activation ling attention,representing space,and providing of brain regions representing other unattended coherence and continuity to conscious visual ex- features of that object (O'Craven,Downing.& perience:(1)attention makes a selection from Kanwisher,2000). object-based representations of space:(2)inde Obiect-based models predict that brain lesions pendent neural mechanisms that operate in parallel could produce an object-based simultanagnosia that orient attention within objects and between objects: (3)the candidate objects on which attention oper- is independent of location.This is precisely the kind of simultanagnosia that was obs ved in patients ates are generated preattentively by early vision in with Balint's syndrome decades before this debate the absence of explicit awareness;and(4)attention is involved in affording explicit(conscious)access erimental work by Humph to the spatial representations needed for goal- and colleagues has shown that simultanagnosia directed action and for binding features of objects. nifest in no atial domains In tw with parietal lobe lesions and poor s atial localiza tion. these e authors observed that pictures extin
visual scenes (Wolpert, 1924). As defined by Wolpert, the term includes, but is more general than, the constriction of attention seen in Bálint’s syndrome. It is seen in conditions other than Bálint’s syndrome and may result from unilateral lesions. Hécaen and de Ajuriaguerra describe the difficulties of one of their patients (case 1) on being offered a light for a cigarette: “[W]hen the flame was offered to him an inch or two away from the cigarette held between his lips, he was unable to se the flame because his eyes were fixed on the cigarette” (Hécaen & de Ajuriaguerra, 1956, p. 374). However, the mechanism underlying simultanagnosia in such cases may be different than that which causes simultanagnosia in Bálint’s syndrome. Unlike in Bálint’s syndrome, simultanagnosia caused by unilateral left temporoparietal lesions appears to be due to a perceptual bottleneck caused by slowing of visual processing as measured by rapid, serial, visual presentation (RSVP) tasks (Kinsbourne & Warrington, 1962, 1963). In contrast, patients with Bálint’s syndrome may be able to recognize a series of individual pictures flashed briefly in an RSVP test (Coslett & Saffran, 1991). Implications of Bálint’s Syndrome for Understanding Visual Cognition Bálint’s syndrome holds valuable lessons for understanding the neural processes involved in controlling attention, representing space, and providing coherence and continuity to conscious visual experience: (1) attention makes a selection from object-based representations of space; (2) independent neural mechanisms that operate in parallel orient attention within objects and between objects; (3) the candidate objects on which attention operates are generated preattentively by early vision in the absence of explicit awareness; and (4) attention is involved in affording explicit (conscious) access to the spatial representations needed for goaldirected action and for binding features of objects. Object- and Space-Based Attention An appreciation of simultanagnosia in Bálint’s syndrome has proven influential in helping to resolve one of the major theoretical controversies in visual attention research. The issue at stake was whether visual attention acts by selecting locations or objects. Work by Michael Posner and others (Posner, 1980; Posner, Snyder, & Davidson, 1980) showed that allocating attention to a location in the visual field enhanced the processing of the visual signals that appeared at the attended location. Object-based models of attention, in contrast, postulate that preattentive processes parse the visual scene to generate candidate objects (more on this later) and that attention then acts by selecting one such object for further processing that can guide goal-directed action. These models are supported by experiments in normal individuals that show better discrimination of two features belonging to the same object than of features belonging to two different objects (Duncan, 1984) and that these object-based effects are independent of the spatial location of their features (Baylis & Driver, 1995; Vecera & Farah, 1994). Physiological recordings have shown that an object-based attentional set can modulate processing in the extrastriate visual cortex (Chelazzi, Duncan, Miller, & Desimone, 1998). Recent neuroimaging studies have confirmed that attentional selection of one of two objects results in activation of brain regions representing other unattended features of that object (O’Craven, Downing, & Kanwisher, 2000). Object-based models predict that brain lesions could produce an object-based simultanagnosia that is independent of location. This is precisely the kind of simultanagnosia that was observed in patients with Bálint’s syndrome decades before this debate was joined by psychologists and physiologists. Moreover, recent experimental work by Humphreys and colleagues has shown that simultanagnosia can be manifest in nonspatial domains. In two patients with parietal lobe lesions and poor spatial localization, these authors observed that pictures extinRobert Rafal 34
Balint's Syndrome 35 guished words and closed shapes extinguished between two parts of the same object.G.K.was open shapes(Humphreys,Romani,Olson,Riddoch, correct on 84%of the trials,whereas in condition 2 Duncan.1994).Thus the object-based attention in which the judgment required comparison of two deficit in this syndrome cannot be attributed simply separate objects,performance was at chance level to the effects of parietal lobe lesions in disrupting (54%. access to spatial representations. Visual Processing Outside of Conscious Neural Representations of Objects in Space Awareness The spatial representations upon which attention The interaction of spatial and object representations ates are det rmined by obi "candidate in det ng the allocation of atter om a d that candidate objects be tion require vided by pr early 9941 s that ed in P of Cmulative c atio with h he olect (see nd. ha posite rates or mechan rating paralle g at pn ys ome ex an o ided ng evic ce to 1号 11 sive proce ssing of visual information outside of awareness objec Balint's syndrome and 2.G.K. Preattentive Representation of Space ondi task was to report whether the upright segments Spatial disorientation is a cardinal feature of were the same or different lengths.For the stimuli Balint's syndrome,and one view of the constriction in condition I.in which the comparison was CONDITION Unequal length Equal length Number correct 67180 Complete U (84%) Separate 10 45/80 (54%) Figure 2.6 Figures used by Cooper and Humphreys (2000)to demonstrate grouping in Balint's syndrome
guished words and closed shapes extinguished open shapes (Humphreys, Romani, Olson, Riddoch, & Duncan, 1994). Thus the object-based attention deficit in this syndrome cannot be attributed simply to the effects of parietal lobe lesions in disrupting access to spatial representations. Neural Representations of Objects in Space The spatial representations upon which attention operates are determined by objects, or “candidate” objects, derived from a grouped array of features by early vision (Vecera & Farah, 1994), and are not simple Cartesian coordinates of empty space centered on the observer (Humphreys, 1998). Humphreys has recently posited that attention operates on spatial representations determined by objects, and that there are separate mechanisms, operating in parallel, for shifting attention within objects and between objects (Humphreys, 1998). Shifting attention within an object implies shifting attention between locations within the object. Figure 2.6 shows stimuli that Cooper and Humphreys (2000) used to study shifts of attention within and between objects in patient G.K. with Bálint’s syndrome. In conditions 1 and 2, G.K.’s task was to report whether the upright segments were the same or different lengths. For the stimuli in condition 1, in which the comparison was between two parts of the same object, G.K. was correct on 84% of the trials, whereas in condition 2 in which the judgment required comparison of two separate objects, performance was at chance level (54%). Visual Processing Outside of Conscious Awareness The interaction of spatial and object representations in determining the allocation of attention requires that candidate objects be provided by preattentive processes that proceed in the absence of awareness. Cumulative observations in patients with hemispatial neglect (see chapter 1) have indeed provided growing evidence that early vision does separate figure from ground, group features, and assign primary axes; it even extracts semantic information that can assign attentional priorities for subsequent processing. Here some examples are considered in which implicit measures of processing in Bálint’s syndrome have provided strong evidence for extensive processing of visual information outside of awareness. Preattentive Representation of Space Spatial disorientation is a cardinal feature of Bálint’s syndrome, and one view of the constriction Balint’s Syndrome 35 Figure 2.6 Figures used by Cooper and Humphreys (2000) to demonstrate grouping in Bálint’s syndrome
Robert Rafal of visual attention posits that it,too,is due to a loss syndrome.Furthermo e,although this infor nation 15) of a neural re sentation of s on which atten is not consciously ssible to the atient.it doe influence the on of obie from the ork of Hu example.Coslett Saffran (1991) et al.(1994).s also nted of words or s briefly to the the and ents ith b P aw only one (Da and B sh reis dman-Hil Hence Grabo c1997 the meaning nage does not el nate repr or whether one or ion,but rather prevents explicit access to perceived nforma an example of hierarchical stimuli in Roberts n et al.(1997)showed that although wh evel and th patient R.M.o (Baylis,Driver,Baylis Rafal,1994)showed patient R.M letter strings spatial Stroop interference effect.That is,although and asked him to report all the letters he could see he could not report whether the word"up"was in Since he could only see one letter at a time,he found the upper or lower visual field,he was,nevertheless. this task ifficult and,with the brief exposure dura- slower to read"up"if it appeared in the lower visual tions used in the experiment.he usually only saw a field than in the upper visual field. few of the letters.However,when the letter string Preattentive Grouping of Features and constituted a word,he was able to report more letters than when it did not.That is.even when the Alignment of Principal Axis patient was naming letters and ignoring the word, the word was processed and helped to bring the con- 1959 stituent letters to his awareness. in the visual are connec ed Othe Attention,Spatial Representation,and Feature recent Humph eys his col Integration:Gluing the World Together that group- collinea ity,surrour I discussed earlier how a single object seen by a rity a are patient is experientially mutable in time.It has no is grouping based on alignment of a principal axis past or future.Any object that moves disappears.In Figure 2.7 shows G.K.'s performance in reporting addition,objects seen in the present can be per- two items:it shows that performance is plexing to the patient,because other objects that when the items are grouped on the basis of bright- the patient does not see.and their features.are ness,collinearity,connectedness,surroundness,and familiarity (Humphreys,1998). such as its color and its shape.are correctly con Preattentive Processing of Meaning of Words ioined because visual attention selects the locat of the obiect and glues together all the feature As is the case in hemispatial neglect. neglecte charino that so e location (Treisman Gelade 1980).For the syndrome wever,all locations are the same.and all the
of visual attention posits that it, too, is due to a loss of a neural representation of space on which attention may act (Friedman-Hill et al., 1995). However, as we have seen from the work of Humphreys et al. (1994), simultanagnosia may also occur for nonspatial information, such as shifting between words and pictures. Moreover, recent observations in patients with both hemispatial neglect (Danziger, Kingstone, & Rafal, 1998) and Bálint’s syndrome (Robertson, Treisman, Friedman-Hill, & Grabowecky, 1997) have shown that parietal damage does not eliminate representations of spatial information, but rather prevents explicit access to this information. Robertson et al. (1997) showed that although patient R.M. could not explicitly report the relative location of two objects, he nevertheless exhibited a spatial Stroop interference effect. That is, although he could not report whether the word “up” was in the upper or lower visual field, he was, nevertheless, slower to read “up” if it appeared in the lower visual field than in the upper visual field. Preattentive Grouping of Features and Alignment of Principal Axis As described earlier, observations by Luria (Luria, 1959) and by Humphreys & Riddoch (1993) have revealed that there is less simultanagnosia when shapes in the visual field are connected. Other recent observations by Humphreys and his colleagues in patient G.K. have confirmed that grouping based on brightness, collinearity, surroundeness, and familiarity also are generated preattentively, as is grouping based on alignment of a principal axis. Figure 2.7 shows G.K.’s performance in reporting two items; it shows that performance is better when the items are grouped on the basis of brightness, collinearity, connectedness, surroundness, and familiarity (Humphreys, 1998). Preattentive Processing of Meaning of Words As is the case in hemispatial neglect, neglected objects do appear to be processed to a high level of semantic classification in patients with Bálint’s syndrome. Furthermore, although this information is not consciously accessible to the patient, it does influence the perception of objects that are seen. For example, Coslett & Saffran (1991) simultaneously presented pairs of words or pictures briefly to their patient, and asked her to read or name them. When the two stimuli were not related semantically, the patient usually saw only one of them, but when they were related, she was more likely to see them both. Hence, both stimuli must have been processed to a semantic level of representation, and the meaning of the words or objects determined whether one or both would be perceived. Words are an example of hierarchical stimuli in which letters are present at the local level and the word at the global level. We (Baylis, Driver, Baylis, & Rafal, 1994) showed patient R.M. letter strings and asked him to report all the letters he could see. Since he could only see one letter at a time, he found this task difficult and, with the brief exposure durations used in the experiment, he usually only saw a few of the letters. However, when the letter string constituted a word, he was able to report more letters than when it did not. That is, even when the patient was naming letters and ignoring the word, the word was processed and helped to bring the constituent letters to his awareness. Attention, Spatial Representation, and Feature Integration: Gluing the World Together I discussed earlier how a single object seen by a patient is experientially mutable in time. It has no past or future. Any object that moves disappears. In addition, objects seen in the present can be perplexing to the patient, because other objects that the patient does not see, and their features, are processed and impinge upon the experience of the attended object. Normally, the features of an object, such as its color and its shape, are correctly conjoined, because visual attention selects the location of the object and glues together all the features sharing that same location (Treisman & Gelade, 1980). For the patient with Bálint’s syndrome, however, all locations are the same, and all the Robert Rafal 36
Balint's Syndrome 32 GROUPING FACTOR TWO ITEM RESPONSES Baseline 7/30 Brightness 18/30 Collinearity 25/30 Connectedness ● 27/30 Surroundedness 28/30 Familiarity 30/30 Figure 2.7 Figures used by Humphreys (1998)to demonstrate grouping in Balint's syndrome. features that impinge on the patient's awareness are Robertson and her colleagues (1997)argued that that obiect. tial on did that feat bindin e red leme owed R.M.pairs a relati and asked hi implicit r.R.M. aw nexcep con ve repo ha Kanw etter that he Lackin oop para to a sp. rep col hich co by his cons cte o report th lor and igno ough h field ble to report exp itly were free loating and cor was colored was neverth a large joined arbitrarily.Since a spatialS op enect was Stroop interferenc effect (i.e he was slower to observed in patient R.M.(see earlier discussion). name the color)when a word had an incongruent
features that impinge on the patient’s awareness are perceptually conjoined into that object. Friedman-Hill et al. (1995) showed R.M. pairs of colored letters and asked him to report the letter he saw and its color. R.M. saw an exceptional number of illusory conjunctions (Treisman & Schmidt, 1982), reporting the color of the letter that he did not see as being the color of the letter that he did report. Lacking access to a spatial representation in which colocated features could be coregistered by his constricted visual attention, visual features throughout the field were free floating and conjoined arbitrarily. Since a spatial Stroop effect was observed in patient R.M. (see earlier discussion), Robertson and her colleagues (1997) argued that spatial information did exist and that feature binding relies on a relatively late stage where implicit spatial information is made explicitly accessible. Subsequent observations in patient R.M. showed, however, that feature binding also occurred implicitly. Wojciulik & Kanwisher (1998) used a modification of a Stroop paradigm in which R.M. was shown two words, one of which was colored, and asked to report the color and ignore the words. Although he was not able to report explicitly which word was colored, there was nevertheless a larger Stroop interference effect (i.e., he was slower to name the color) when a word had an incongruent Balint’s Syndrome 37 Figure 2.7 Figures used by Humphreys (1998) to demonstrate grouping in Bálint’s syndrome