Opinion

The Split-Brain Phenomenon

Revisited: A Single Conscious
Agent with Split Perception
Yair Pinto,1,2,* Edward H.F de Haan,1,2,3 and
Victor A.F. Lamme1,2,3
The split-brain phenomenon is caused by the surgical severing of the corpus
Trends
callosum, the main route of communication between the cerebral hemispheres.
Five hallmarks characterize the split-
The classical view of this syndrome asserts that conscious unity is abolished. brain syndrome: a response  visual
The left hemisphere consciously experiences and functions independently of field interaction, strong hemispheric
specialization, confabulations after
the right hemisphere. This view is a cornerstone of current consciousness left-hand actions, split attention, and
research. In this review, we first discuss the evidence for the classical view. the inability to compare stimuli across
We then propose an alternative, the ‘conscious unity, split perception’ model. the midline.

This model asserts that a split brain produces one conscious agent who These hallmarks underlie the classical
experiences two parallel, unintegrated streams of information. In addition to notion that split brain implies split con-
sciousness. This notion suggests that
changing our view of the split-brain phenomenon, this new model also poses a
massive interhemispheric communica-
serious challenge for current dominant theories of consciousness. tion is necessary for conscious unity.

Closer examination challenges the

The Five Hallmarks of the Split-Brain Syndrome classical notion. Either the hallmark
In humans, nearly all communication between the cerebral hemispheres occurs via the corpus also occurs in healthy adults or the
hallmark does not hold up for all
callosum [1–3]. In split-brain patients (callosotomized; see Glossary), the corpus callosum
split-brain patients.
is surgically severed, normally at an adult age, to alleviate otherwise intractable seizures (Box 1).
Thus, in split-brain patients, communication between the left and the right cerebral hemisphere A re-evaluation of the split-brain data
is almost completely abolished. Although these patients behave normally and report to feel suggests a new model that might bet-
ter account for the data. This model
unchanged after the operation [4–6], research has revealed a multitude of marked, and
asserts that a split-brain patient is one
sometimes dramatic, changes (Figure 1, Key Figure; see callosal agenesis for comparison). conscious agent with unintegrated
visual perception.
One task has been particularly useful in documenting the cognitive changes observed in split-
This new model challenges prominent
brain patients [6–10]. In this task, a visual stimulus is either presented to the left visual field or
theories of consciousness, since it
the right visual field. This is ensured by monitoring eye fixations and movements, and by implies that massive communication
presenting stimuli for less than 0.15 s (the minimum amount of time needed to initiate and is not needed for conscious unity.
execute an eye movement). The reason for this set up is that although both eyes project
information to both cerebral hemispheres, all visual information to the left of fixation (i.e., the left
visual field) is exclusively processed by the right cerebral hemisphere, while all visual information 1
Department of Psychology, University
to the right of fixation (the right visual field) is processed solely by the left hemisphere. Another of Amsterdam, Amsterdam, The
key aspect of this task is the response type. The patient either reacts with the left hand, the right Netherlands
2
hand, or verbally. The idea behind this is that the right hemisphere controls the left side of the Amsterdam Brain & Cognition (ABC)
Center, University of Amsterdam,
body, including the left hand, and the left hemisphere controls the right hand and verbal Amsterdam, The Netherlands
responses. Thus, this task controls both which hemisphere receives input and which hemi- 3
These authors contributed equally to
sphere produces output. this manuscript

The prototypical split-brain patient shows five related phenomena on variations of this task. *Correspondence:
Arguably, these five aspects encompass the main differences between split-brain patients and [email protected] (Y. Pinto).

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 http://dx.doi.org/10.1016/j.tics.2017.09.003 835
© 2017 Elsevier Ltd. All rights reserved.
 Box 1. Midline Connections
Glossary
In split-brain patients, the corpus callosum is surgically removed after age 12. The corpus callosum is by far the largest of
Callosal agenesis: person born
the commissures – white matter tracts that connect homologous structures on both sides of the central nervous system
without a corpus callosum. These
– and possesses a complex architecture (e.g., [117,118]). A large part of the corpus callosum, extending from the genu
subjects show few of the
to the posterior part, connects the prefrontal cortices. Fibers from the parietal lobes cross mainly in dorsal areas of the
characteristics of the split-brain
splenium and isthmus, while the temporal lobes are largely connected via the posterior and ventral regions of the corpus
syndrome.
callosum. The medial cortical surface is largely connected via the dorsal corpus callosum, while the fibers from the
Callosotomized: another term to
ventral regions of the brain cross ventrally (e.g., [119,120]). The visual cortices are mainly connected via the splenium
describe a split-brain patient.
[121]. With respect to subcortical structures, there is evidence that the claustrum connects to the contralateral prefrontal
Gaze contingent presentation: in
cortex, precentral gyrus, and postcentral gyrus and claustrum via the body of the corpus callosum [122].
this procedure, gaze fixation is
determined through eye tracking.
The anterior commissure entails connections between the orbitofrontal, temporal, parietal, and occipital lobes [123], Based on this, the visual stimulus is
and even the insular cortices. In terms of subcortical structures, the anterior commissure connects the olfactory bulbs, presented relative to fixation (e.g., 3
the septal area, the amygdalae, and the overlying entorhinal cortices [124]. In addition, a small number of fibers visual degrees to the left or the right
connecting both claustra pass through the anterior commissure [122]. In some split-brain patients this commissure is of fixation). Subsequently, the
removed as well. presentation time is either kept too
short for the subject to move his/her
The posterior commissure connects the precentral and postcentral gyri, the superior parietal region in the left hemi- eyes to the stimulus, or the stimulus
sphere to the temporal region, and lateral occipital and superior parietal regions of the right hemisphere [119]. The moves in sync with the gaze of the
subcortical connections that run through the posterior commissure originate in the thalamic, superior colliculus and the subject.
habenular nuclei. Response  visual field
interaction: there are three different
Other smaller commissures include the hippocampal commissure (connecting the subicular and parahippocampal response types (left hand, right hand,
cortices [124]), the commissure of Probst (connecting the dorsal nucleus of the lateral lemniscus and the inferior and verbally) and two visual fields
colliculus), the commissure of the inferior colliculi (connecting the two inferior colliculi), the commissure of the superior (left and right). If a subject is more
colliculi (connecting the two superior colliculi), the habenular commissure (which connects the habenular nuclei), the accurate in responding to the left
middle commissure (connecting the thalamus), and the anterior and posterior cerebellar commissures (connecting the visual field with the left hand than
two cerebellar hemispheres). Finally, many fibers decussate in lower brain structures, such as the pons. Examples are with the right hand/verbally or more
the white matter tracts from the two hemispheres of the cerebellum to the cortical hemispheres (e.g., [125]). In general, accurate in responding to the right
the smaller commissures and decussations are intact in split-brain patients. visual field with the right hand/
verbally than with the left hand, then
a response  visual field interaction
exists. However, if accuracy does
healthy adults. The first, and most salient, observation in split-brain patients is the not depend on a combination of
response  visual field interaction [1,5,6,8,9,11]. When a stimulus is presented to the left response type and visual field, then
this interaction is absent. If, for
visual field, the patient can only respond adequately with his/her left hand, and vice versa for the
instance, a subject is better at
right field and hand. The second aspect concerns the hemispheric specialization, with each of responding with the left hand, but
the two hemispheres being better at certain tasks [8,10,12–14]. The third aspect focusses on irrespective of whether stimuli appear
post hoc confabulations after actions with the left hand [1,15]. The fourth characteristic in the left or the right visual field,
then this interaction is absent.
concerns the observation that each hemisphere may have its own independent focus of
Similarly, if performance is superior
attention [16–19]. Finally, there is abundant evidence that shows that split-brain patients for stimuli in the right visual field, but
are incapable of comparing stimuli across the visual midline [10,20–23]. irrespective of response type, then
the interaction is also absent.
Split and unified consciousness:
Altogether these five observations have led to what we dub the classical models of split-brain
split consciousness indicates the
patients. There are two primary classical models. The first model revolves around the notion existence of two independent
that only the left hemisphere gives rise to consciousness, while the right hemisphere only conscious agents, whereas unified
processes information in an unconscious manner. The right hemisphere may prime the left consciousness (or conscious unity)
implies the existence of only one
hemisphere toward certain behavior, but this will only affect consciousness after it has been conscious agent.
molded and interpreted by the left hemisphere. This is the so-called partial consciousness Split-brain patient: patients in
model [1,11]. The second classical model posits that in a split-brain patient each hemisphere whom the corpus callosum is
has its own consciousness, independent of the other hemisphere [5,6,24]. Thus, according to surgically removed. This is generally
done to alleviate otherwise
this ‘split consciousness’ model, a split-brain patient houses two independent conscious intractable seizures.
agents. This model has been concisely argued by Sperry [24] who wrote that the two hemi- Visual field: all stimuli appearing to
spheres acted as if they were ‘two separate conscious entities or minds running in parallel in the the left of where a subject fixates are
considered to be part of the left
same cranium, each with its own sensations, perceptions, cognitive processes’ (p. 318).
visual field. Everything to the right of
where the subject fixates is part of
By contrast, we will argue that despite their prominence, these classical models face serious the right visual field.
challenges. This is because three of the five hallmarks (hemispheric specialization, post hoc

836 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

Key Figure
The Five Hallmarks of the Split-Brain Syndrome

I: Response x visual field II: Hemispheric specializaon III: Le hemisphere

interacon confabulaon

What did you

Tim
ls
draw that for?

tai

e
De
Maths Visual I, uhmm, saw a nice
+ house on the way +

Ca
Nothing here
age

usa
gu

o
Lan

IV: Split aenon V: Cannot compare across

midline

Same?

+ +
??

Figure 1. The classical view of split-brain patients asserts that conscious unity is disrupted in this syndrome. The evidence
for this view comes from five hallmarks. First, a marked response type  visual field interaction occurs in split-brain patients
[1,5,6,8,9,11]. They can only respond accurately to stimuli in the right visual field with the right hand or verbally, and to
stimuli in the left visual field with the left hand. Therefore, when a stimulus appears in the left visual field, the patient verbally
reports that he/she saw nothing, yet draws the image with his/her left hand. This supports the notion that each hemisphere
controls half the body, and consciously perceives half the visual field. The second hallmark is extreme hemispheric
specialization [8,10,12–14,36–45]. The left hemisphere is, among other things, better at language, maths, and detailed
processing. The right hemisphere is better at visuospatial tasks, time perception, and causal inferencing. This again
suggests that each hemisphere operates independently of the other, and thereby creates consciousness autonomously.
The third striking phenomenon is that split-brain patients confabulate wildly when asked to explain actions of their left hand
(controlled by the mute right hemisphere) [1,15]. The notion here is that the left hemisphere creates an independent
conscious agent, who is unaware of why the right hemisphere chooses its actions. Therefore, this agent resorts to ad hoc
confabulations. Fourth, in split-brain patients, each hemisphere seems to have its own focus of attention [16–19]. Since
attention and consciousness are thought to be tightly linked [64–67], this again supports the classical notion that
consciousness is not unified in split-brain patients. Fifth, split-brain patients cannot compare stimuli across the midline
[10,20–23,85–87]. This makes sense if two independent conscious agents each view half of the visual field, and cannot
communicate their perceptions to each other.

confabulations, and split attention) also exist in healthy adults with unified consciousness.
Thus, these hallmarks cannot constitute proof for disturbances in conscious unity. Moreover,
one hallmark (inability to compare across the midline) also fits the more modest explanation that
visual processing is unintegrated. In general, a more extreme explanation (destroyed conscious
unity) should only be preferred if simpler explanations do not suffice. Finally, the strongest proof
for a breakdown of unified consciousness, the response  visual field interaction, does not hold
for all split-brain patients. We posit a new model of the split-brain syndrome, which claims that
both hemispheres give rise to a single conscious agent, and discuss its wider implications.

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 837

Hallmark 1: Response  Visual Field Interaction
The most striking effect of a severed corpus callosum is a remarkable response type  visual
field interaction. When a stimulus is presented to the right visual field, the patient responds
essentially in a normal manner. However, when a stimulus is presented to the left visual field,
the patient verbally indicates that he/she saw nothing, yet with his/her left hand he/she
indicates that he/she did see the stimulus. When probed, the patient cannot explain the action
of his/her left hand [1,5,6,8,9,11,15]. This is a key data point for the classical models, since it
dramatically shows that normal consciousness and control are disrupted. In healthy adults,
where both hemispheres give rise to one conscious agent, a response type  visual field
interaction is completely absent. Therefore, this finding is perhaps the most well-known split-
brain finding, and is highlighted in textbooks and reviews [25–27]. Both classical models
explain this finding along the same lines. Concordant with human anatomy, the right visual
field is solely processed by the left hemisphere, which controls language and the right hand
[28–33]. Thus, when a stimulus appears in the right visual field, the patient will produce a
normal verbal response. However, when stimuli are presented to the left visual field, they are
processed by the mute right hemisphere, which controls the left hand. Thus, in such a case,
the left hemisphere will verbally, and truthfully, indicate that it saw nothing. Yet, the right
hemisphere will indicate that it ‘did’ see a stimulus, and will make this clear by its control of the
left hand. However, the models disagree on whether the action initiated by the right hemi-
sphere is based on an automatic response (the partial consciousness model) or on conscious
control (the split consciousness model).

On closer examination, the response  visual field interaction appears less than beyond
dispute. First, Sperry [6] himself, when describing the response  visual field interaction, notes
in the last paragraph that ‘Although the general picture has continued to hold up in the main as
described . . . striking modifications and even outright exceptions can be found among the
small group of patients examined to date’ (p 733), suggesting that the response  visual field
interaction may be less absolute than commonly assumed. Similarly, other studies [7,16] also
cast doubt on the generality of a response  visual field interaction in split-brain patients. In one
study [7] split-brain patients viewed chimeric faces, with one half of the face presented to the left
visual field, and the other to the right visual field. The patients then either indicated the name of
the face they had just seen (when they responded verbally), or pointed to one of three possible
faces (when they responded manually). When responding verbally, they overwhelmingly named
the face that had appeared in the right visual field. However, when they selected a face by
pointing, they overwhelmingly selected the face presented to the left visual field. Crucially, this
was even the case when they responded with the right hand. Another study [16] investigated
attention in split-brain patients, and controlled for the role of response type (left hand or right
hand). Interestingly, although they found an effect of visual field, response type did not play a
significant role.

More recently, we [10] performed a quantitative study into the interaction, using sophisticated
fixation control with an eye tracker, a substantial number of trials in each condition, forced-
choice responding, and a large number of different stimuli. The response type (left hand, right
hand, or verbally) was varied systematically. We found, in two split-brain patients, that
although visual field played a large role in most tasks, a response type  visual field interaction
was never observed. This result held across all tasks (detection, localization, orientation
determination, labeling, and visual matching), and all tested types of stimuli (isoluminant dots,
simple shapes, oriented rectangles, objects). For instance, when the subject indicated
changes in orientation of a stimulus, performance was superior for stimuli in the left visual
field, even when he/she responded verbally, or with the right hand. The result also held for
high confidence trials, suggesting that the lack of interaction cannot be explained by implicit
processing [34,35]. Thus, although the textbook claim of an absolute response  visual field

838 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

interaction would be strong evidence for the classical models, closer examination reveals
that, at least in some patients, this interaction is absent. Consequently, this hallmark cannot
be taken as evidence that conscious unity necessarily breaks down when the corpus
callosum is removed.

Why are the results we obtained so different from the results obtained in some earlier
research? One possibility is that split-brain patients show strong individual differences [6].
Another intriguing possibility is that the difference is due to a change in methodology. Note
that the earlier claims of a strong response  visual field interaction are almost solely
based on qualitative reports (Table 1, Key Table). In our recent research, we not only
ensured a quantitative approach, but also ensured complete clarity of the task. For
instance, when the patient verbally labeled objects, he/she initially tended to indicate that
‘he/she saw nothing’, when an image was presented to the left visual field. However, after
clarifying the instructions and adding catch trials where nothing was presented, these
responses entirely disappeared. Rather, now the patient only indicated that he/she saw
nothing on the catch trials, and (sometimes) reported an inability to name stimuli when they
appeared in the left visual field. Thus, it is also possible that with our new, rigorous
methodology virtually no split-brain patients would show a response  visual field interac-
tion on accuracy.

Hallmark 2: Hemispheric Specialization

Another key hallmark is that split-brain patients show pronounced hemispheric specialization.
This is in line with the notion that each hemisphere operates largely independently of the other,
and thus that conscious unity is disturbed. The general procedure for determining hemispheric
differences in split-brain patients is by presenting an image/task either to the left visual field
(which is processed by the right hemisphere) or to the right visual field (processed by the left
hemisphere). Therefore, when a patient is better at naming objects when they are presented to
the right visual field, this leads to the conclusion that the left hemisphere is better at naming. This
methodology has revealed an abundance of hemispheric specialization in split-brain patients.
For instance, the left hemisphere is superior in language production and verbal labeling of
images [8,10,12–14], solving mathematical problems [36], recognizing local details [37], and
self-recognition [38]. The right hemisphere is better at visual-spatial tasks [10,39], causal
inference [40], temporal discrimination [41], object-recognition based on fragments [42–44],
and detecting statistical regularities in visual scenes [45].

Although the data here are unequivocal, they also do not provide conclusive evidence for the
classical models, since hemispheric specialization is also observed in healthy adults. Using a
similar procedure as in split-brain patients, the left hemisphere has been found to be superior in
abstract category processing [46,47], recognizing words [48], processing high spatial fre-
quency [49–52], local processing [53,54], and recognizing positive emotions [55,56]. The right
hemisphere shows superior performance for stimulus-specific visual processing [46,47], rec-
ognizing faces [48], processing low spatial frequency [49–52], global processing [53,54], and
recognizing negative emotions [55,56]. Thus, also in a healthy adult human brain it seems that
different types of processes run in parallel in the different cerebral hemispheres, without
apparently disturbing conscious unity.

Hallmark 3: Post Hoc Confabulation

An intriguing split-brain phenomenon is the observed post hoc confabulations to explain one’s
own behavior [1,15]. In one example [15], a split-brain patient was shown a picture of a bell
tower in his left visual field. He then indicated which picture he had just seen, by pointing to one
of four pictures with his left hand. When asked why he chose the picture, the patient indicated
that he ‘must have heard a bell ringing on [his] way into the lab’.

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 839

Key Table
Table 1. Overview of the Best Known Split-Brain Patients on the Hallmarks of the Split-Brain
Syndromea
Initi- Patient info R  VF, QLc R  VF, QNd Cannot Can compare, Cannot Split attention
alsb compare, V2f compare, Tg
V1e

W.J. Age of operation: 48 Yes [8] Yes [23] Yes [8] Yes [8]
Extent: corpus
callosum [8]

L.B. Age of operation: Yes [6] No [7,128–130] Yes [86] Yes [87,88,133] Yes, visual search [18,19]
13, corpus callosum Mixed [131,132] Mixed [132] No, negative priming [134]
[126,127]

N.G. Age of operation: 30 Yes [6] No [7,128–130] No [86] Yes [88] No [23] Yes [18]
Extent: corpus Mixed [23] No, negative priming [134]
callosum [126,127]

A.A. Age of operation: 13 Yes [6] No [7,128,135] Yes [86] Yes [88]
Extent: corpus
callosum [126,127]

R.Y. Age of operation: Yes [6] Yes [128] Yes [86]

43, corpus callosum
[126,127]

C.C. Age of operation: 13 Yes [6] Mixed [7,128]

Extent: corpus
callosum [126,127]

M.E. Age: 19 No [136]

Extent: corpus
callosum [136]

J.W. Age: 26 Yes [1] No [16,138,139] Yes [140] Mixed [132] Yes, visual search [18,19]
Extent: corpus Mixed [132] No [141] No [142] Yes, object-based attention [16,17]
callosum [137] Yes, guided search [143]
Yes, object-based inhibition [144]
No, attentional blink [81]
No, spatial cueing [83,84]
No, object-based facilitation [144]

D.D.C. Age: 19 No [10,147] Yes [10] Yes [89] No [148]

Extent: corpus (Y. Pinto et al.,
callosum, most of unpublished)
the anterior
commissure
[145,146]

D.D.V. Age: 22 No [10] Yes [142] Mixed [148]

Extent: corpus
callosum [145,146]

a
Hemispheric specialization is not listed, since all patients show this strongly. Confabulation is not listed, since there are almost no studies into this topic.
b
Patient V.P. is not listed, because her callosotomy turned out to be incomplete.
c
R  VF, QL: response  visual field interaction, but only qualitative report.
d
R  VF, QN: response  visual field interaction, but (also) quantitative report. Interaction only refers to the ability to accurately respond to stimuli in left visual field,
verbally and with the right hand, and to stimuli in right visual field with the left hand, not to reaction time differences for incongruent responses (since these also occur for
healthy subjects).
e
V1 refers to cross hemifield comparisons that are generally thought to be impossible for split-brain patients (numbers, orientation, shapes, etc.).
f
V2 refers to cross hemifield comparisons that should be possible (apparent motion, gestalt/gist perception).
g
T refers to interhemispheric comparisons based on tactile information.

840 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

Both the classical models ascribe the post hoc confabulations to the breakdown of conscious
unity. The initial information is projected to the right hemisphere, which controls the left hand.
However, the verbal explanation is provided by the left hemisphere. Thus, the explanation is
given by a conscious agent who did not initiate or execute the action. Therefore, the left
hemisphere confabulates an answer. The classical models only differ on whether the actions by
the right hemisphere are automatic (partial consciousness model) or consciously controlled
(split consciousness model).

Crucially, like hemispheric specialization, post hoc confabulation is not unique to split-brain
patients. First, also in healthy adults, behavior can often be caused by factors that the subject is
unaware of. For instance, when people communicate they seem to engage in automatic
mimicry [57,58], where they mirror the behavior of the other without being aware of doing
so. A more dramatic example is that when people are shopping for wine, the background music
seems to affect their choices (when listening to French music, people are more inclined to buy
French wine [59,60]). Second, when probed about the reasons for their choices, even when
these choices were consciously made, subjects are not immune to wild confabulations. The
phenomenon of ‘choice blindness’ clearly illustrates this [61–63]. In choice blindness, subjects
are asked to choose from two options, for instance, which of two people they find most
attractive. After the subject has made his/her choice, the experimenter, on some trials, swaps
the choice outcome, and now presents the rejected option to the subject as the one that was
chosen. If the subject is subsequently asked why he/she had chosen this option (the one he/
she actually rejected) over the other (the option he/she actually chose), the most frequent
response is an impromptu confabulation. This has also been found for gustatory, aesthetic, and
moral choices. Thus, also in healthy adults (with unified consciousness) behavior can be elicited
by factors unknown to the subject. Moreover, healthy subjects may confabulate to explain their
own behavior.

Hallmark 4: Split Attention

It is often argued that attention and consciousness are closely related ([64–67], but see also
[68–71]). Therefore, key support for the classical models has been drawn from studies that
suggest that attention is split in split-brain patients (Figure 2). Several studies suggest that, in
split-brain patients, object-based and space-based attention are situated in different hemi-
spheres [16,17], implying independent attentional centers. Moreover, when split-brain patients
search for a target among distractors, search efficiency doubles when the distractors are
distributed across two, rather than one, visual hemifields [18,19]. This suggests that each
hemisphere autonomously scans half of the visual field.

However, these findings do not conclusively show a disturbance of conscious unity for two
reasons. First, again similar phenomena are observed in healthy adults. In multiple object
tracking [72–75] subjects attentively follow a subset of dots. When the dots are presented to
both visual fields, subjects are able to track two times as many dots as when all stimuli are
presented to one visual field [76,77], suggesting that each hemisphere has a fixed capacity and
tracks the moving dots independently of the other. Moreover, when a tracked dot crosses the
midline, and thus passes from one visual field to the other, the electroencephalogram signal
indicates that one hemisphere passes the relevant information of the moving dot to the other
hemisphere [78]. This confirms the notion that each hemisphere tracks information indepen-
dently of the other, and only shares information when necessary.

Second, in addition to ‘split attention’ occurring in healthy subjects, attention can also be unified
in split-brain patients (Figure 2). Normally, when subjects have to detect two targets in a rapid
serial visual presentation, for instance, two letters in a stream of numbers, performance is
lowest when the time between the first and the second target is 200–500 ms. This is called the

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 841

 Split aenon
(A) Visual search (B) Object-based aenon

+ +
Within object
Unilateral Bilateral
Cue
Unilateral search slope

Reacon me cost in seconds(s)

Bilateral search slope Betw objects
Paents 49.2 Controls
Reacon me in seconds(s)

0.1
1100
44.7
Within
Betw
39.4
27.8
0.06

600 0.02
2 4 8 2 4 8 LVF RVF
Set size Set size

Unified aenon
(C) Precue (D) Aenonal blink
Invalid Invalid M
K
interhemifield intrahemifield P
T2 2
C
e

+ + T1
Tim

4
A
T
V B
+
+ X + X E P

70 Interhemifield
Reacon me in seconds(s)

Interhemifield Intrahemifield
1.9 Same stream
Intrahemifield
% accuracy

40
1.5

1.1
10
Valid Neutral Invalid Lag 2 Lag 8

Figure 2. Split and Unified Attention in Split-Brain Patients. Evidence has been found both for and against split
attention in split-brain patients. When split-brain patients search through a bilateral display (A) they are two times as
efficient as searching through a unilateral display [18,19]. This effect is not observed in controls, who are equally efficient in
both cases. This suggests that in split-brain patients each hemisphere autonomously scans half of the visual field.
Moreover, in split-brain patients, reaction time costs for invalidly cued items (B) are higher when the invalidly cued location
is on a different object. However, this is only the case for objects appearing in the right visual field, suggesting that only the
attentional center in the left hemisphere employs object-based attention [16]. By contrast, behavior of split-brain patients in
a Posner-cueing [116] paradigm (C) reveals evidence for unified attention [83]. A valid attentional precue accelerates
reaction times, while an invalid one slows them. This pattern holds both within and across hemispheres, suggesting that
(Figure legend continued on the bottom of the next page.)

842 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

attentional blink [79,80]. Moreover, when there are multiple rapid serial visual presentation
streams, this attentional blink transfers between streams. Therefore, if the first target appears in
stream 1, and the second target in stream 2, detection of the second target is still lowest when
time between the first and second target is between 200 and 500 ms. Crucially, split-brain
patients show a similar transfer of attentional blink [81,82]. When the first target is presented in
one visual field, split-brain patients still show a normal attentional blink for the second target,
even if it appears in the other visual field. Thus, when it comes to the attentional blink (unlike in
visual search) attention is unified. Moreover, attentional unity in split-brain patients also seems
preserved for attentional precueing. When a precue highlights a location in the other visual field,
split-brain patients still benefit from it [83]. In addition, when a target can appear in two
locations, one in each visual field, a singular precue is effective, regardless of whether the
precue appears in the same or the other visual field. However, two simultaneous precues, one
per visual field, are ineffective [84].

Hallmark 5: Inability to Compare Stimuli across the Midline

Split-brain patients seem incapable of comparing stimuli across the midline, irrespective of the
type of stimulus (digits, simple shapes, rectangles, or images) [10,20–23,85–87] (see Figure 3).
In other words, when one stimulus is presented to the left visual field and the other to the right
visual field, the patient cannot accurately indicate whether both stimuli are the same, although
he/she can do so when both stimuli are presented within one visual field. This is consistent with
the notion that each hemisphere independently perceives the contralateral visual field (the right
hemisphere possibly perceiving the left visual field unconsciously).

Again, the data are more mixed than generally thought. Although there are indeed many
examples of split-brain patients who are incapable of comparing stimuli across the midline,
prominent examples can also be found of patients who can compare stimuli across the midline.
For instance, one study found that although some split-brain patients could not compare stimuli
across the midline, others could do so well above chance [86] (Figure 3). Moreover, under certain
circumstances nearly all tested split-brain patients seem able to compare stimuli across visual
fields. For instance, in one experiment two tilted lines were presented with a gap in between
them. The lines were positioned in such a way that extending them across the gap would either
cause the lines to coincide or to run in parallel. When split-brain patients had to indicate which
was the case, they could do so highly accurately, even when both line segments were located in
different visual fields [87] (Y. Pinto et al., unpublished). In addition, when a split-brain patient
viewed a visual field consisting of many circles, he/she could accurately indicate on which side
the circle with the largest average radius appeared (Y. Pinto et al., unpublished). A final example
of visual integration across the midline involves apparent motion. When two dots were presented
in succession at a short distance (2 to 14 visual degrees; see gaze contingent presentation),
split-brain patients were able to accurately indicate whether the dots created apparent motion, or
that they were presented simultaneously or with delays too long to create apparent motion.
Critically, they were able to do so even when one dot appeared in the left visual field, and the
other in the right visual field [87–89]. Note that although this capability is found in nearly all tested
patients, individual differences may play a role here, since one split-brain patient was not able to
perceive apparent motion across the midline [90]. Thus, although split-brain patients generally
cannot compare stimuli across the midline, they may be able to do so in some special cases,
perhaps involving gestalt principles [91] or statistical/gist processing [92–94].

unified attentional resources are employed. The attentional blink of split-brain patients also suggests unified attention (D).
When controls view multiple streams consisting of items presented in rapid succession, they have a ‘universal’ attentional
blink. That is, when the first target (in this example a number among letters) is detected, the second target is less well
detected at lag 2 in all other streams. Split-brain patients also show a universal attentional blink, even when the first and the
second target appear in different visual fields [81,82]. Abbreviation: Betw, between.

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 843

 + 5 + 2 +
+

+ + +

Figure 3. Visual Integration across Hemispheres without a Corpus Callosum. Investigations into visual integra-
tion in split-brain patients reveal an equivocal picture. Most patients cannot perform simple comparisons across the
midline, such as comparing simple shapes, numbers, orientation, or pictures [10,20–23,85–87]. Yet, other types of visual
integration do seem possible. Split-brain patients can generally indicate whether two interrupted lines would or would not
coincide across the midline [87] (Y. Pinto et al., unpublished), whether an object produces apparent motion across the
midline [87–89], and which visual field contains the largest circle (Y. Pinto et al., unpublished).

A New Model: One Conscious Agent with Unintegrated Visual Perception

Critically considering all five characteristics of the split-brain phenomenon, our first conclusion is
that none of the data provide compelling proof for the central tenet of the classical models, that
is, consciousness is split in split-brain patients. Hemispheric dominance and post hoc con-
fabulations cannot be proof for split consciousness, since they also occur in healthy adults with
unified consciousness. Split attention cannot be proof, since this also sometimes occurs in
healthy adults, and moreover attention is sometimes unified in split-brain patients. A respon-
se  visual field interaction would be a very strong proof, since it never occurs in healthy adults,
and is exactly what you would expect to observe if consciousness is split. However, this
characteristic is not universally true for split-brain patients.

Moreover, not only is conclusive evidence for disturbances in conscious unity lacking, there is
also positive evidence for preserved conscious unity in split-brain patients. First, a breakdown
of conscious unity seems incompatible with the observation that split-brain patients often
behave and feel normally [4–6,10]. It seems implausible that both the patient himself/herself and
the people in close contact with the patient fail to observe that after the surgical removal of the
corpus callosum the original agent only perceives half the visual field and controls only half the
body. Even when probed explicitly, the patient still insists that he/she controls his/her entire
body and can see perfectly fine at both sides of fixation [10].

Second, controlled laboratory tests confirm this self-assessment. The absence of a response
type  visual field interaction in (some) split-brain patients seems incompatible with split
consciousness. Imagine one conscious agent controlling language and the right hand (the
left hemisphere) and another (conscious or unconscious) agent controlling the left hand.
Furthermore, the former agent only perceives the right visual field, and the latter the left visual
field. This model seems to logically imply a response type  visual field interaction, where
performance for the left visual field is better when the patient responds with the left hand than
with the right hand/verbally (and vice versa).

Thus, our new model asserts that consciousness is unified in split-brain patients. This can explain
why these patients feel normal and behave normally. They have not become half blind and half
paralyzed after the surgery (and forced to share their brain with another conscious being). Rather,
they are still one conscious agent in control of their entire body, and thus feel mostly unchanged.
This one agent experiences the entire visual field, and controls his/her entire body. Therefore, it is
trivially easy for this agent to respond to stimuli anywhere in the visual field, with any response
type, even with the left hand to stimuli in the right visual field and vice versa. In other words, the

844 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

preserved conscious unity that our model hypothesizes for split-brain patients can explain why
they are mostly normal in their behavior and show no response  visual field interaction.

However, this cannot be the whole story, since split-brain patients do show marked differences
from healthy adults. Many comparisons across the midline that are trivially easy for healthy
adults prove impossible for split-brain patients. Moreover, hemispheric specialization is much
more pronounced in these patients than in healthy adults. We argue that these differences can
be explained by something less extreme than split consciousness, namely, by unintegrated
perception. That is, we posit that in split-brain patients visual perception is (largely) unintegrated
across visual fields. Therefore, they cannot make comparisons across the midline. Moreover,
this unintegrated perception can also explain the increased hemispheric specialization. In
healthy adults, hemispheric specialization occurs if the contralateral hemisphere contributes
more to visual processing than the ipsilateral one. In split-brain patients, we posit that visual
information is not shared, and thus that the contralateral hemisphere does all the visual
processing, leading to maximal hemispheric specialization.

In short, the ‘conscious unity, split perception’ model (Figure 4) asserts that the split-brain
patient is one conscious agent, in whom visual perception remains unintegrated across
hemifields. What could this be like from a first-person perspective? We speculate that this
can be thought of as watching an ‘out-of-sync’ movie. When you watch such a movie, your
conscious unity is preserved, in the sense that you are still one conscious agent. Yet, your

(A)

Cognion Motor +
(e.g., comparison) language

Split visual Divided Divided Speech

informaon cognion control agents

Cognion Motor
(e.g., comparison)

(B)

Cognion
(e.g., comparison)

Central
Split visual Divided Speech
control
informaon cognion
agent

Cognion
(e.g., comparison)

Figure 4. Different Models of the Split-Brain Syndrome. The classic and alternative ways to explain the set of split-
brain findings discussed in this paper. Classically (A), split-brain patients are considered to have their visual information,
cognition (e.g., the ability to compare objects), and their output control mechanisms split, each sitting in separate
hemispheres, with language predominantly processed and produced by the left hemisphere. Alternatively (B), and better
fitting the findings on hemifield  response type interaction, both visual information and cognition might indeed be split
across the midline (so that objects cannot be compared across the midline). Yet each visual half field and cognition module
can be monitored by a single central agent, which simultaneously controls both hands and speech (as in normal subjects).

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 845

perception is split, since you experience a visual and an auditory stream, and you are unable to
integrate these two streams. The strength of our new model is that it can explain both the
normalities (because consciousness is unified) and the abnormalities (because perception is
split) of the split-brain syndrome. Notice that this model makes a surprising prediction. Although
the split-brain patient is not able to compare visual stimuli across the midline when they are
presented simultaneously (because there is no automatic integration of information), he/she
should be able to do so when the stimuli are presented with a considerable delay in between
them. Since the patient experiences all the relevant information, he/she should be able to use a
conscious strategy to resolve the task. Thus, whereas for normal people simultaneous com-
parisons are easier than sequential ones, this should be reversed for split-brain patients.

Implications for Leading Theories of Consciousness

The classical models of the split-brain syndrome have had an enormous impact on our view of
consciousness. These models assert that without a corpus callosum, and thus without massive
communication between both cerebral hemispheres, conscious unity is destroyed. The idea
that unified consciousness requires massive, direct communication has served as a corner-
stone in consciousness research. Thus, our new model, which suggests that unified con-
sciousness may arise without massive communication, challenges currently leading theories on
consciousness, such as the integrated information theory (IIT [95–97]), the global workspace
(GW) theory [64,98,99], and the recurrent processing (RP) theory [100–102]. According to IIT,
consciousness arises when a system has a rich representation repertoire and its subsystems
are strongly interconnected. However, when the integration within a subsystem is larger than
the connection between subsystems, consciousness will arise as a function of the subsystem
rather than of the system as a whole. In split-brain patients, intrahemispheric connectivity
exceeds interhemispheric connectivity, and thus each hemisphere should give rise to an
independent conscious agent [96].

GW theory makes a similar prediction as IIT. According to this theory, the cerebral hemispheres
house a ‘global workspace’ [64,98,99]. This GW can be thought of as the ‘headquarters’ of the
cortex, which receives information from and projects to many cortical modules. Only informa-
tion processed by the GW reaches consciousness. Crucially, when the two cerebral hemi-
spheres are prevented from exchanging information, they can no longer give rise to one
integrated GW [98]. Consequently, either only one hemisphere has a GW or both hemispheres
create their own GW. Thus, GW theory implies that in split-brain patients partial or split
consciousness arises.

RP theory asserts that consciousness can arise through local RP between cortical modules,
even in the absence of global or integrative cortical processes [100–102]. However, such local
processing only leads to phenomenal consciousness that is otherwise inaccessible, and
unreportable. Thus, for unified ‘reportable’ consciousness, strong integration between the
hemispheres is still needed. Yet, RP theory could be more compatible with our new model than
IIT and GW theory. It could also be argued that, according to RP theory, a more basic,
nonverbalizable type of consciousness is not dependent on interhemispheric integration [101].
Thus, whether the newly proposed model supports or challenges RP theory depends on the
specifics. If consciousness is unified in all respects, including reporting capacities, then it does
challenge. However, if a split-brain patient is one conscious agent, but with split reporting
capacities, then it may be compatible with RP theory.

Subcortical or Functional Unity as the Basis of Unified Consciousness

Interestingly, two less prominent notions within consciousness research may be able to explain
preserved conscious unity in split-brain patients. First, subcortical areas are still unified in split-
brain patients (Box 1). Thus, if consciousness is critically linked to subcortical processing, such

846 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

as thalamic processing [103,104], then conscious unity may be unaffected in the split-brain
syndrome. For instance, if corticothalamic loops drive consciousness [104], then it is possible
that the specific contents of consciousness depend on cortical activity, but that the unity of
consciousness is rooted in the thalamic structures involved in these loops.

Second, it is possible that conscious unity is strongly related to functional unity [66,105].
Functional unity means that different parts of a system operate mostly in conjunction, rather
than independently. Note that functional unity can arise when subsystems are connected to a
common source, even when they are not directly linked. For instance, according to the
sensorimotor account of consciousness [66], conscious experiences are driven by the func-
tional structure of processed information, rather than by the specific cortical implementation.
Moreover, if the different cortical systems in a split-brain patient are driven by common input
and produce conjoined output, then functional unity could still arise in the cortex, even without
direct links between both cerebral hemispheres. Supporting the idea that functional unity is
preserved in split-brain patients is the finding that interhemispheric neural activity in resting state
is highly synchronized in split-brain patients, even as synchronized as resting state activity in
healthy subjects [106]. If consciousness is indeed mostly dependent on functional structure,
then preserved functional unity could underlie conscious unity in the split-brain syndrome.

Now, what about the neuroanatomical underpinning of this central agent? Broadly speaking,
two explanations seem most likely. One possibility, in line with the subcortical theory of
consciousness, is that the unification of consciousness relies on certain key hardware require-
ments. Perhaps unified consciousness remains intact when a minimal number of axonal
connections between subcortical structures exist, or certain key parts are directly connected.
Another possibility is more in line with the functional notion of consciousness. According to this
explanation, consciousness remains unified if a minimal amount of synchronization between
subsystems is preserved, regardless of whether this synchrony is driven by direct or indirect
connections. The former explanation suggests that consciousness will split if more (subcortical)
connections between the hemispheres are cut, even if synchrony between the hemispheres is
preserved. The latter account predicts that even without any direct connections, conscious-
ness can remain unified if synchrony remains intact.

Concluding Remarks and Future Perspectives

The status of conscious unity in the split-brain syndrome is crucial for our broader understand-
ing of consciousness. The classical view of the syndrome, which asserts that unified con-
sciousness is destroyed, supports the currently leading theories of consciousness. However, if
conscious unity is preserved, then subcortical or functionally unifying processes likely play an
important role in consciousness. The classical view is based on five defining features of the
split-brain syndrome. Split-brain patients show strongly pronounced hemispheric specializa-
tion, cannot verbally explain their left-hand movements, have independent centers of attention,
cannot compare stimuli across the midline and, most importantly, can only respond accurately
to stimuli in the left visual field with the left hand, and to stimuli in the right visual field with the
right hand/verbally (the response  visual field interaction).

We argue that the classical view may not hold for several reasons. First, some of the defining
features also occur in healthy adults with unified consciousness (hemispheric specialization,
inability to explain own actions, and split attention). Second, the most convincing argument
against unified consciousness in split-brain patients (the response  visual field interaction)
does not hold for all split-brain patients. Third, in the absence of any convincing proof against
split consciousness, unified consciousness should be the default position. Both the patients
and the people nearest to them claim that consciousness is still unified in the patient. Moreover,
their everyday behavior confirms this. Thus, the claim of destroyed conscious unity is

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 847

extraordinary, and requires extraordinary evidence. Past research seemed to provide this Outstanding Questions
extraordinary proof through the response  visual field interaction, but recent research has Can a split-brain patient overcome his/
invalidated this proof. her inability to compare stimuli across
visual fields? In general, a split-brain
patient cannot compare visual infor-
Further progress on the issue of unified consciousness with a split brain could come not only mation across the midline. However,
from studies with split-brain patients, but also through studying ‘temporary split brain’ (a if the patient is really one conscious
agent, who is aware of two indepen-
temporary suspension of communication between neural modules in a healthy subject), so
dent streams of information, can the
one can experience the syndrome from a first-person perspective. Different manipulations of comparison be made through con-
attention may result in a temporary split brain. As mentioned, in multiple object tracking each scious effort? For instance, if stimuli
hemisphere seems to track dots independently. If further behavioral and neural research are presented across the midline
sequentially, with a long delay in
confirms this, then multiple object tracking or multiple identity tracking [73,74,107] tasks
between, can the patient develop an
may provide a tool for creating a transient split-brain state. Moreover, during inattentional effortful strategy to accomplish accu-
blindness [108–110], unattended information seems to be processed only locally [111,112]. rate comparisons?
Thus, strongly focusing attention may also create a temporary split brain, where attended and
unattended information is processed independently. A window into the consciousness of the Where does the ability to integrate
visual information across the midline
unattended information is available during a short time frame after disappearance of the stimuli,
fall apart? The patient is able to accu-
during the so-called fragile memory stage [113–115]. Recent work [68] suggests that unat- rately indicate whether two lines would
tended information in fragile memory elicits normal consciousness. Only after information is coincide or run in parallel if they would
transferred from fragile to working memory (a different stage of visual short-term memory be extended across the midline. How-
ever, he/she cannot determine
[68,114,115]) does unattended information become unconscious [68]. Future research may whether two bars, appearing in two
unveil whether this unattended information indeed gives rise to normal conscious experiences visual fields, have the same orientation.
while eliciting only local neural processing, virtually insulated from other cortical processing. Can he/she only visually integrate lines,
Several other questions are currently unanswered (see Outstanding Questions). For instance, if or can bars also be integrated? Do the
stimuli need to have the same degree
a split-brain patient is really one conscious agent, can the individual overcome the lack of visual of rotation? Are the differences in abil-
integration by conscious effort? Why are some forms of visual integration possible, and others ity driven by task requirements (indi-
not? What are the neural mechanisms underlying unified control in split-brain patients? cate similarity or indicate continuity)
rather than differences in low-level
properties?
Acknowledgments
This work was supported by ERC Advanced grant FAB4V (#339374), and an ERC Advanced grant to V.A.F.L.
What are the neural principles under-
lying unified control in a split-brain
References patient? When the patient responds
1. Gazzaniga, M.S. (2000) Cerebral specialization and interhemi- 12. Gazzaniga, M.S. et al. (1984) Profiles of right hemisphere lan- with the left hand to stimuli in the right
spheric communication: does the corpus callosum enable the guage and speech following brain bisection. Brain Lang. 22,
visual field, is this movement controlled
human condition? Brain 123, 1293–1326 206–220
by the ipsilateral or the contralateral
2. Innocenti, G.M. (1986) General organization of callosal 13. Gazzaniga, M.S. et al. (1963) Laterality effects in somesthesis
following cerebral commissurotomy in man. Neuropsychologia
cerebral hemisphere? If the patient
connections in the cerebral cortex. In Sensory-Motor Areas
and Aspects of Cortical Connectivity, pp. 291–353, 1, 209–215 responds by performing a mental
Springer 14. Gazzaniga, M.S. (1975) Review of the split brain. J. Neurol. 209, action (e.g., think of swimming when
3. Wahl, M. et al. (2007) Human motor corpus callosum: topog- 75–79 the answer is A, think of arithmetics
raphy, somatotopy, and link between microstructure and func- 15. Corballis, P.M. (2003) Visuospatial processing and the right- when the answer is B), can the correct
tion. J. Neurosci. 27, 12132–12138 hemisphere interpreter. Brain Cogn. 53, 171–176 response then be decoded from both
4. Bogen, J.E. et al. (1965) Cerebral commissurotomy: a second 16. Egly, R. et al. (1994) Covert orienting in the split brain reveals hemispheres, or only from the contra-
case report. JAMA 194, 1328–1329 hemispheric specialization for object-based attention. Psychol. lateral hemisphere?
5. Sperry, R. (1984) Consciousness, personal identity and the Sci. 5, 380–383
divided brain. Neuropsychologia 22, 661–673 17. Kingstone, A. (2016) Covert orienting in the split brain: right
hemisphere specialization for object-based attention. Laterality
What would it be like to be a split-brain
6. Sperry, R.W. (1968) Hemisphere deconnection and unity in
conscious awareness. Am. Psychol. 23, 723 21, 732–744 patient? Third-person observations
18. Luck, S.J. et al. (1994) Independent attentional scanning in the can only tell us so much about unified
7. Levy, J. et al. (1972) Perception of bilateral chimeric figures
following hemispheric deconnexion. Brain 95, 61–78 separated hemispheres of split-brain patients. J. Cogn. Neuro- consciousness in others; the ultimate
8. Gazzaniga, M.S. et al. (1962) Some functional effects of sec-
sci. 6, 84–91 test would come from a first-person
tioning the cerebral commissures in man. Proc. Natl. Acad. Sci. 19. Luck, S.J. et al. (1989) Independent hemispheric attentional perspective. Is it possible to create a
U. S. A. 48, 1765–1769 systems mediate visual-search in split-brain patients. Nature temporary split brain through atten-
342, 543–545
9. Gazzaniga, M.S. (1967) The split brain in man. Sci. Am. 217, 24– tional, or other, manipulations?
29 20. Fendrich, R. and Gazzaniga, M.S. (1989) Evidence of foveal
splitting in a commissurotomy patient. Neuropsychologia 27,
10. Pinto, Y. et al. (2017) Split brain: divided perception but undi-
273–281
vided consciousness. Brain 140, 1231–1237
21. Corballis, M.C. and Sergent, J. (1992) Judgments about numer-
11. Wolford, G. et al. (2000) The left hemisphere’s role in hypothesis
osity by a commissurotomized subject. Neuropsychologia 30,
formation. J. Neurosci. 20, 1–4
865–876

848 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

22. Corballis, M.C. (1994) Can commissurotomized subjects com- 49. Jonsson, J.E. and Hellige, J.B. (1986) Lateralized effects of
pare digits between the visual fields? Neuropsychologia 32, blurring: a test of the visual spatial frequency model of cerebral
1475–1486 hemisphere asymmetry. Neuropsychologia 24, 351–362
23. Gazzaniga, M.S. et al. (1965) Observations on visual perception 50. Sergent, J. (1983) The effects of sensory limitations on hemi-
after disconnexion of the cerebral hemispheres in man. Brain 88, spheric processing. Can. J. Psychol. 37, 345–366
221–236 51. Sergent, J. (1983) Role of the input in visual hemispheric asym-
24. Sperry, R.W. (1966) Mental unity following surgical disconnec- metries. Psychol. Bull. 93, 481–512
tion of the cerebral hemispheres. Harvey Lect. 62, 52. Michimata, C. and Hellige, J.B. (1987) Effects of blurring and
293–323 stimulus size on the lateralized processing of nonverbal stimuli.
25. Gazzaniga, M.S. et al. (1998) The problem of consciousness. In Neuropsychologia 25, 397–407
Cognitive Neuroscience – The Biology of the Mind, pp. 542– 53. Fink, G.R. et al. (1996) Where in the brain does visual attention
543, W.W. Norton & Company select the forest and the trees? Nature 382, 626–628
26. Gray, P. (2002) The nervous system. In Psychology, pp. 160– 54. Kimchi, R. and Merhav, I. (1991) Hemispheric processing of
161, Worth Publishers global form, local form, and texture. Acta Psychol. 76, 133–147
27. Wolman, D. (2012) A tale of two halves. Nature 483, 260–263 55. Davidson, R.J. et al. (1987) Ratings of emotion in faces are
28. Cowey, A. (1979) Cortical maps and visual perception the Grind- influenced by the visual field to which stimuli are presented.
ley memorial lecture. Q. J. Exp. Psychol. 31, 1–17 Brain Cogn. 6, 403–411
29. Sakata, H. and Taira, M. (1994) Parietal control of hand action. 56. Reuter-Lorenz, P. and Davidson, R.J. (1981) Differential contri-
Curr. Opin. Neurobiol. 4, 847–856 butions of the two cerebral hemispheres to the perception of
30. Penfield, W. and Boldrey, E. (1937) Somatic motor and sensory happy and sad faces. Neuropsychologia 19, 609–613
representation in the cerebral cortex of man as studied by 57. Lakin, J.L. et al. (2008) I am too just like you: nonconscious
electrical stimulation. Brain 60, 389–443 mimicry as an automatic behavioral response to social exclu-
31. Cantalupo, C. and Hopkins, W.D. (2001) Asymmetric Broca’s sion. Psychol. Sci. 19, 816–822
area in great apes. Nature 414, 505–505 58. Dimberg, U. et al. (2002) Facial reactions to emotional stimuli:
32. Vigneau, M. et al. (2006) Meta-analyzing left hemisphere lan- automatically controlled emotional responses. Cogn. Emot. 16,
guage areas: phonology, semantics, and sentence processing. 449–471
Neuroimage 30, 1414–1432 59. North, A.C. et al. (1999) The influence of in-store music on wine
33. Ojemann, G. et al. (1989) Cortical language localization in left, selections. J. Appl. Psychol. 84, 271–276
dominant hemisphere: an electrical stimulation mapping inves- 60. North, A.C. et al. (1997) In-store music affects product choice.
tigation in 117 patients. J. Neurosurg. 71, 316–326 Nature 390, 132–132
34. Lau, H. and Rosenthal, D. (2011) Empirical support for higher- 61. Hall, L. et al. (2012) Lifting the veil of morality: choice blindness
order theories of conscious awareness. Trends Cogn. Sci. 15, and attitude reversals on a self-transforming survey. PLoS One
365–373 7, e45457
35. Rosenthal, D. (2000) Consciousness and metacognition. In 62. Johansson, P. et al. (2005) Failure to detect mismatches
Metarepresentation: Proceedings of the Tenth Vancouver Cog- between intention and outcome in a simple decision task. Sci-
nitive Science Conference (Sperber, D., ed.), pp. 265–295, ence 310, 116–119
Oxford University Press 63. Hall, L. et al. (2010) Magic at the marketplace: choice blindness
36. Funnell, M.G. et al. (2007) The calculating hemispheres: studies for the taste of jam and the smell of tea. Cognition 117, 54–61
of a split-brain patient. Neuropsychologia 45, 2378–2386 64. Dehaene, S. and Naccache, L. (2001) Towards a cognitive
37. Martin, M. (1979) Hemispheric specialization for local and global neuroscience of consciousness: basic evidence and a work-
processing. Neuropsychologia 17, 33–40 space framework. Cognition 79, 1–37
38. Turk, D.J. et al. (2002) Mike or me? Self-recognition in a split- 65. Posner, M.I. (1994) Attention: the mechanisms of conscious-
brain patient. Nat. Neurosci. 5, 841–842 ness. Proc. Natl. Acad. Sci. U. S. A. 91, 7398–7403
39. Corballis, P. et al. (2002) Hemispheric asymmetries for simple 66. O’Regan, J.K. and Noë, A. (2001) A sensorimotor account of
visual judgments in the split brain. Neuropsychologia 40, vision and visual consciousness. Behav. Brain Sci. 24, 939–973
401–410 67. Cohen, M.A. et al. (2012) The attentional requirements of con-
40. Roser, M.E. et al. (2005) Dissociating processes supporting sciousness. Trends Cogn. Sci. 16, 411–417
causal perception and causal inference in the brain. Neuropsy- 68. Pinto, Y. et al. (2017) Conscious visual memory with minimal
chology 19, 591 attention. J. Exp. Psychol. Gen. 146, 214–226
41. Funnell, M.G. et al. (2003) Temporal discrimination in the split 69. Vandenbroucke, A.R. et al. (2012) Non-attended representa-
brain. Brain Cogn. 53, 218–222 tions are perceptual rather than unconscious in nature. PLoS
42. Nebes, R.D. (1972) Dominance of the minor hemisphere in One 7, e50042
commissurotomized man on a test of figural unification. Brain 70. Lamme, V.A. (2003) Why visual attention and awareness are
95, 633–638 different. Trends Cogn. Sci. 7, 12–18
43. Nebes, R.D. (1971) Superiority of the minor hemisphere in 71. Koch, C. and Tsuchiya, N. (2007) Attention and consciousness:
commissurotomized man for the perception of part-whole rela- two distinct brain processes. Trends Cogn. Sci. 11, 16–22
tions. Cortex 7, 333–349
72. Howe, P.D. et al. (2010) Distinguishing between parallel and
44. Nebes, R.D. (1974) Hemispheric specialization in commissuro- serial accounts of multiple object tracking. J. Vis. 10, 11
tomized man. Psychol. Bull. 81, 1–14
73. Pinto, Y. et al. (2010) The more often you see an object, the
45. Roser, M.E. et al. (2011) Right hemisphere dominance in visual easier it becomes to track it. J. Vis. 10, 4
statistical learning. J. Cogn. Neurosci. 23, 1088–1099
74. Pinto, Y. et al. (2012) Tracking moving identities: after attending
46. Marsolek, C.J. (1999) Dissociable neural subsystems underlie the right location, the identity does not come for free. PLoS One
abstract and specific object recognition. Psychol. Sci. 10, 111– 7, e42929
118
75. Pylyshyn, Z.W. and Storm, R.W. (1988) Tracking multiple inde-
47. Laeng, B. et al. (2003) Do separate processes identify pendent targets: evidence for a parallel tracking mechanism.
objects as exemplars versus members of basic-level catego- Spat. Vis. 3, 179–197
ries? Evidence from hemispheric specialization. Brain Cogn. 53,
76. Alvarez, G.A. and Cavanagh, P. (2005) Independent resources
15–27
for attentional tracking in the left and right visual hemifields.
48. Dundas, E.M. et al. (2013) The joint development of hemispheric Psychol. Sci. 16, 637–643
lateralization for words and faces. J. Exp. Psychol. Gen. 142, 77. Cavanagh, P. and Alvarez, G.A. (2005) Tracking multiple targets
348–358 with multifocal attention. Trends Cogn. Sci. 9, 349–354

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 849

78. Drew, T. et al. (2014) A soft handoff of attention between cere- 106. Uddin, L.Q. et al. (2008) Residual functional connectivity in the
bral hemispheres. Curr. Biol. 24, 1133–1137 split-brain revealed with resting-state functional MRI. Neurore-
79. Raymond, J.E. et al. (1992) Temporary suppression of visual port 19, 703–709
processing in an RSVP task: an attentional blink? J. Exp. Psy- 107. Horowitz, T.S. et al. (2007) Tracking unique objects. Atten.
chol. 18, 849–860 Percept. Psychophys. 69, 172–184
80. Shapiro, K.L. et al. (1997) The attentional blink. Trends Cogn. 108. Rock, I. et al. (1992) Perception without attention: results of a
Sci. 1, 291–296 new method. Cogn. Psychol. 24, 502–534
81. Giesbrecht, B. and Kingstone, A. (2004) Right hemisphere 109. Mack, A. and Rock, I. (1998) Inattentional Blindness, MIT Press
involvement in the attentional blink: evidence from a split-brain 110. Simons, D.J. and Chabris, C.F. (1999) Gorillas in our midst:
patient. Brain Cogn. 55, 303–306 sustained inattentional blindness for dynamic events. Percep-
82. Ptito, A. et al. (2009) The attentional blink within and across the tion 28, 1059–1074
hemispheres: evidence from a patient with a complete section of 111. Tsuchiya, N. et al. (2015) No-report paradigms: extracting the
the corpus callosum. Biol. Psychol. 82, 64–69 true neural correlates of consciousness. Trends Cogn. Sci. 19,
83. Holtzman, J.D. et al. (1984) Spatial orientation following com- 757–770
misural section. In Varieties of Attention (Parasuram, R. and 112. Sasai, S. et al. (2016) Functional split brain in a driving/listening
Davies, D.R., eds), pp. 375–394, Academic Press paradigm. Proc. Natl. Acad. Sci. U. S. A. 113, 14444–14449
84. Holtzman, J.D. et al. (1981) Dissociation of spatial information for 113. Landman, R. et al. (2003) Large capacity storage of integrated
stimulus localization and the control of attention. Brain 104, objects before change blindness. Vision Res. 43, 149–164
861–872
114. Sligte, I.G. et al. (2008) Are there multiple visual short-term
85. Sergent, J. (1986) Subcortical coordination of hemisphere activ- memory stores? PLoS One 3, e1699
ity in commissurotomized patients. Brain 109, 357–369
115. Pinto, Y. et al. (2013) Fragile visual short-term memory is an
86. Johnson, L.E. (1984) Bilateral visual cross-integration by human object-based and location-specific store. Psychon. Bull. Rev.
forebrain commissurotomy subjects. Neuropsychologia 22, 20, 732–739
167–175
116. Posner, M.I. (1980) Orienting of attention. Q. J. Exp. Psychol. 32,
87. Corballis, M.C. (1995) Visual integration in the split brain. Neuro- 3–25
psychologia 33, 937–959
117. Bastiani, M. et al. (2012) Human cortical connectome recon-
88. Ramachandran, V. et al. (1986) Perception of apparent motion struction from diffusion weighted MRI: the effect of tractography
by commissurotomy patients. Nature 320, 358–359 algorithm. Neuroimage 62, 1732–1749
89. Knapen, T. et al. (2012) Perception of apparent motion in a split- 118. Sporns, O. et al. (2005) The human connectome: a structural
brain observer. J. Vis. 12, 548–548 description of the human brain. PLoS Comput. Biol. 1, e42
90. Gazzaniga, M.S. (1987) Perceptual and attentional processes 119. Pannek, K. et al. (2010) An automated strategy for the delinea-
following callosal section in humans. Neuropsychologia 25, tion and parcellation of commissural pathways suitable for clini-
119–133 cal populations utilising high angular resolution diffusion imaging
91. Spelke, E.S. (1990) Principles of object perception. Cogn. Sci. tractography. Neuroimage 50, 1044–1053
14, 29–56 120. Park, H. et al. (2008) Corpus callosal connection mapping using
92. Chong, S.C. and Treisman, A. (2005) Statistical processing: cortical gray matter parcellation and DT-MRI. Hum. Brain Mapp.
computing the average size in perceptual groups. Vision Res. 29, 503–516
45, 891–900 121. Wei, P. et al. (2017) Connection between bilateral temporal
93. Greene, M.R. and Oliva, A. (2009) Recognition of natural scenes regions: tractography using human connectome data and dif-
from global properties: seeing the forest without representing fusion spectrum imaging. J. Clin. Neurosci. 39, 103–108
the trees. Cogn. Psychol. 58, 137–176 122. Arrigo, A. et al. (2015) Inter-hemispheric claustral connections in
94. Wolfe, J.M. et al. (2011) Visual search in scenes involves human brain: a constrained spherical deconvolution-based
selective and nonselective pathways. Trends Cogn. Sci. 15, study. Clin. Neuroradiol. http://dx.doi.org/10.1007/s00062-
77–84 015-0492-x December 22, 2015
95. Tononi, G. (2004) An information integration theory of con- 123. Patel, M.D. et al. (2010) Distribution and fibre field similarity
sciousness. BMC Neuroscience 5, 42 mapping of the human anterior commissure fibres by diffusion
96. Tononi, G. (2005) Consciousness, information integration, and tensor imaging. MAGMA 23, 399–408
the brain. Prog. Brain Res. 150, 109–126 124. Raybaud, C. (2010) The corpus callosum, the other great fore-
97. Tononi, G. and Koch, C. (2015) Consciousness: here, there and brain commissures, and the septum pellucidum: anatomy,
everywhere? Philos. Trans. R. Soc. Lond. B Biol. Sci. 370, development, and malformation. Neuroradiology 52, 447–477
20140167 125. Jbabdi, S. et al. (2015) Measuring macroscopic brain connec-
98. Baars, B.J. (2005) Global workspace theory of consciousness: tions in vivo. Nat. Neurosci. 18, 1546–1555
toward a cognitive neuroscience of human experience. Prog. 126. Levy, J. and Trevarthen, C. (1977) Perceptual, semantic and
Brain Res. 150, 45–53 phonetic aspects of elementary language processes in split-
99. Baars, B.J. (1988) A Cognitive Theory of Consciousness, Cam- brain patients. Brain 100, 105–118
bridge University Press 127. Bogen, J.E. et al. (1988) Completeness of callosotomy shown
100. Block, N. (2007) Consciousness, accessibility, and the mesh by magnetic resonance imaging in the long term. Arch. Neurol.
between psychology and neuroscience. Behav. Brain Sci. 30, 45, 1203–1205
481–499 128. Levy, J. and Trevarthen, C. (1981) Color-matching, color-nam-
101. Lamme, V.A. (2006) Towards a true neural stance on conscious- ing and color-memory in split-brain patients. Neuropsychologia
ness. Trends Cogn. Sci. 10, 494–501 19, 523–541

102. Lamme, V.A. and Roelfsema, P.R. (2000) The distinct modes of 129. Zaidel, E. (1998) Stereognosis in the chronic split brain: hemi-
vision offered by feedforward and recurrent processing. Trends spheric differences, ipsilateral control and sensory integration
Neurosci. 23, 571–579 across the midline. Neuropsychologia 36, 1033–1047

103. Churchland, P.S. (2005) A neurophilosophical slant on con- 130. Iacoboni, M. and Zaidel, E. (1995) Channels of the corpus
sciousness research. Prog. Brain Res. 149, 285–293 callosum: evidence from simple reaction times to lateralized
flashes in the normal and the split brain. Brain 118, 779–788
104. Min, B. (2010) A thalamic reticular networking model of con-
sciousness. Theor. Biol. Med. Model. 7, 10 131. Forster, B. and Corballis, M.C. (2000) Interhemispheric transfer
of colour and shape information in the presence and absence of
105. Dennett, D.C. (1991) Consciousness Explained, Little, Brown
the corpus callosum. Neuropsychologia 38, 32–45
and Co.

850 Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11

132. Forster, B. et al. (2000) Effect of luminance on successiveness 141. Gazzaniga, M.S. et al. (1996) Collaboration between the hemi-
discrimination in the absence of the corpus callosum. Neuro- spheres of a callosotomy patient: emerging right hemisphere
psychologia 38, 441–450 speech and the left hemisphere interpreter. Brain 119,
133. Naikar, N. and Corballis, M.C. (1996) Perception of apparent 1255–1262
motion across the retinal midline following commissurotomy. 142. Corballis, M.C. et al. (2004) Hemispheric integration and differ-
Neuropsychologia 34, 297–309 ences in perception of a line-motion illusion in the divided brain.
134. Lambert, A.J. (1993) Attentional interaction in the split-brain: Neuropsychologia 42, 1852–1857
evidence from negative priming. Neuropsychologia 31, 313–324 143. Kingstone, A. et al. (1995) Guided visual search is a left-hemi-
135. Mooshagian, E. et al. (2009) Spatial attention and interhemi- sphere process in split-brain patients. Psychol. Sci. 6,
spheric visuomotor integration in the absence of the corpus 118–121
callosum. Neuropsychologia 47, 933–937 144. Tipper, S.P. et al. (1997) Object-based facilitation and inhibition
136. Aglioti, S. et al. (1993) Hemispheric control of unilateral and from visual orienting in the human split-brain. J. Exp. Psychol.
bilateral responses to lateralized light stimuli after callosotomy Hum. Percept. Perform. 23, 1522–1532
and in callosal agenesis. Exp. Brain Res. 95, 151–165 145. Pizzini, F. et al. (2010) Diffusion tensor tracking of callosal fibers
137. Gazzaniga, M.S. et al. (1985) MRI assessment of human callosal several years after callosotomy. Brain Res. 1312, 10–17
surgery with neuropsychological correlates. Neurology 35, 146. Corballis, M.C. et al. (2010) Mirror-image discrimination and
1763–1766 reversal in the disconnected hemispheres. Neuropsychologia
138. Reuter-Lorenz, P. et al. (1995) Fate of neglected targets: a 48, 1664–1669
chronometric analysis of redundant target effects in the bisected 147. Savazzi, S. et al. (2007) Interhemispheric transfer following
brain. J. Exp. Psychol. Hum. Percept. Perform. 21, 211–230 callosotomy in humans: role of the superior colliculus. Neuro-
139. Handy, T.C. et al. (2003) Cortical and subcortical contributions psychologia 45, 2417–2427
to the representation of temporal information. Neuropsychologia 148. Fabri, M. et al. (2005) Contribution of posterior corpus callosum
41, 1461–1473 to the interhemispheric transfer of tactile information. Cogn.
140. Grabowecky, M. and Kingstone, A. (2004) Can semantic infor- Brain Res. 24, 73–80
mation be transferred between hemispheres in the split-brain?
Brain Cogn. 55, 310–313

Trends in Cognitive Sciences, November 2017, Vol. 21, No. 11 851