Abstract: Recently, the discussion regarding the consequences of cutting the corpus callosum (“split-brain”) has regained momentum (Corballis, Corballis, Berlucchi, & Marzi, Brain, 141(6), e46, 2018; Pinto et al., Brain, 140(5), 1231–1237, 2017a; Pinto, Lamme, & de Haan, Brain, 140(11), e68, 2017; Volz & Gazzaniga, Brain, 140(7), 2051–2060, 2017; Volz, Hillyard, Miller, & Gazzaniga, Brain, 141(3), e15, 2018). This collective review paper aims to summarize the empirical common ground, to delineate the different interpretations, and to identify the remaining questions. In short, callosotomy leads to a broad breakdown of functional integration ranging from perception to attention. However, the breakdown is not absolute as several processes, such as action control, seem to remain unified. Disagreement exists about the responsible mechanisms for this remaining unity. The main issue concerns the first-person perspective of a split-brain patient. Does a split-brain harbor a split consciousness or is consciousness unified? The current consensus is that the body of evidence is insufficient to answer this question, and different suggestions are made with respect to how future studies might address this paucity. In addition, it is suggested that the answers might not be a simple yes or no but that intermediate conceptualizations need to be considered.
[...]
Thus, it seems that in split-brain patients perceptual processing is largely split, yet response selection and action control appear to be unified under certain conditions. This, by itself, does not prove whether a split-brain houses one or two conscious agents. One explanation could be that the split-brain houses two agents, each having their own experiences, who synchronize their behavioral output through various means. Another possible explanation is that a split-brain houses one agent who experiences an unintegrated stream of information who controls the entire body, comparable to watching a movie where sight and sound are out-of-sync. At any rate, these findings challenge the previously mentioned classic split-brain description, which is still found in reviews and text books (Gray, 2002; Wolman, 2012). In this classic characterization the patient indicates that they saw nothing when a stimulus appeared in the left visual field. Yet, to their own verbal surprise, the left hand correctly draws the stimulus. The aforementioned examples of unity in action control suggests that these effects may depend on the type and complexity of the response that is required.
Interpretations
There are three, not-mutually exclusive, hypotheses concerning the mechanisms involved in, seemingly, preserved unity in the split-brain. The first notion is that information is transferred subcortically. The second idea is that ipsilateral motor control underlies unity in action control. The third idea claims that information transfer is based on varies forms of inter-hemispheric collaboration, including subtle behavioral cues. The first proposal (Corballis Corballis, Berlucchi, & Marzi, 2018; de Haan et al., 2019; Pinto, Lamme, & de Haan, 2017b; Pinto et al., 2017a; Savazzi et al., 2007; Mancuso, Uddin, Nani, Costa, & Cauda, 2019) suggests that the multitude of subcortical connections that are spared during surgery are responsible for the transfer of information. As was initially pointed out by Trevarthen (1968) and Trevarthen and Sperry (1973) and recently stressed by Pinto, de Haan, and Lamme (2017a) and Corballis et al. (2018), there are many commissures (white matter tracts that connect homologous structures on both sides of the central nervous system) and decussations (bundles that connect different structures on both sides) that link nuclei that are known to be involved in perceptual processing. The importance of these commisural connections for transferring visual information in split-brain patients has been highlighted by Trevarthen and Sperry (1973). Moreover, the role of these connections in a split-brain has recently been demonstrated by bilateral fMRI activations in the first somatosensory cortex, after unilateral stimulation of trunk midline touch receptors (Fabri et al., 2006) and in the second somatic sensory area after unilateral stimulation of hand pain receptors (Fabri, Polonara, Quattrini, & Salvolini, 2002). Uddin and colleagues used low-frequency BOLD fMRI resting state imaging to investigate functional connectivity between the two hemispheres in a patient in whom all major cerebral commissures had been cut (Uddin et al., 2008). Compared to control subjects, the patient’s interhemispheric correlation scores fell within the normal range for at least two symmetrical regions. In addition, Nomi and colleagues suggested that split-brain patients might rely particularly on dorsal and ventral pontine decussations of the cortico-cerebellar interhemispheric pathways as evidenced by increased fractional anisotropy (FA) on diffusion weighted imaging (Nomi, Marshall, Zaidel, Biswal, Castellanos, Dick, Uddin & Mooshagian, 2019). Interhemispheric exchange of information also seems to occur in the domain of taste sensitivity, activation of primary gustatory cortex in the fronto-parietal operculum was reported in both hemispheres after unilateral gustatory stimulation of the tongue receptors (Mascioli, Berlucchi, Pierpaoli, Salvolini, Barbaresi, Fabri, & Polonara, 2015). Note that patients may differ with respect to how many of these connections have been cut, and this might also explain some of the individual variance among patients. Moreover, in all patients subcortical structures remain intact. For instance, the superior colliculus is known to integrate visual information from both hemispheres and project information to both hemispheres (Meredith & Stein, 1986; Comoli et al., 2003). Such structures may support attentional networks, and may enable the right hemisphere to attend to the entire visual field. In turn, attentional unity could help in unifying cognitive and motor control, which may subserve ipsilateral motor control.
The second point concerns the ipsilateral innervation of the arms. Manual action is not strictly lateralized, and the proximal (but not the distal) parts of the arm are controlled bilaterally, although the ipsilateral contribution remains undetermined. This could explain why split-brain patients may respond equally well with both hands in certain experimental conditions (Corballis, 1995; Gazzaniga, Bogen, & Sperry, 1967; Pinto, de Haan, & Lamme, 2017a). First, there is substantial evidence that bilateral cortical activations can be observed during unilateral limb movements in healthy subjects. In addition, ipsilesional motor problems in arm control have been observed in patients with unilateral cortical injuries, and finally there is evidence from electrocorticography with implanted electrodes for localization of epileptic foci showing similar spatial and spectral encoding of contralateral and ipsilateral limb kinematics (Bundy, Szrama, Pahwa, & Leuthardt, 2018). While these observations argue convincingly for a role in action control by the ipsilateral hemisphere, they do not prove that a hemisphere on it’s own can purposefully control the movements of the ipsilateral hand. Thus, the role of ipsilateral arm-hand control in explaining split-brain findings is currently not settled.
The third hypothesis argues that in addition to whatever direct neural communication may exist between the hemispheres, they may inform one another via strategic cross-cueing processes (Volz & Gazzaniga, 2017; Volz et al., 2018). The split-brain patients underwent surgery many years prior to testing, and the separated perceptual systems have had ample time to learn how to compensate for the lack of commissural connections. For example, subtle cues may be given by minimal movements of the eyes or facial muscles, which might not even be visible to an external observer but are capable of encoding, for example, the location of a stimulus for the hemisphere that did not “see” it. A cross-cueing mechanism might also allow one hemisphere to convey to the other which one of a limited set of known items had been shown (Gazzaniga & Hillyard, 1971; Gazzaniga, 2013).
Finally, it is possible to entertain combinations of the different explanations. For instance, it is conceivable that in the subacute phase following split-brain surgery the hemispheres are ineffective in communicating with each other. During this initial phase, phenomena such as an “alien hand” - that is a hand moving outside conscious control of the (verbal) person - may be present. In the ensuing period, the patients may have learned to utilize the information that is exchanged via subcortical connections, ipsilateral motor control or cross-cueing to coordinate the processing of the two hemispheres. In such a way, the patient may counteract some of the effects of losing the corpus callosum.
What do We Need to Know?
This paper aims to contribute to the agenda for the next decade of split-brain research. Full split-brain surgery is rare these days, and it is important that we try to answer the central questions while these patients are still available for study. In order to examine the variations between patients it would be useful to test as many of the available patients as possible with the same tests.
One important goal is to map out precisely how much functionality and information is still integrated across hemispheres in the split-brain, and what the underlying principles are. For instance, in some cases the two hemispheres seem to carry out sensory-motor tasks, such as visual search, independently from one another (Arguin et al., 2000; Franz, Eliassen, Ivry, & Gazzaniga, 1996; Hazeltine, Weinstein, & Ivry, 2008; Luck, Hillyard, Mangun, & Gazzaniga, 1994; Luck et al., 1989), while in other cases functions such as attentional blink, or attentional cueing, seem to be integrated across hemispheres (Giesbrecht & Kingstone, 2004; Holtzman, Volpe, & Gazzaniga, 1984; Holtzman, Sidtis, Volpe, Wilson, & Gazzaniga, 1981; Pashler et al., 1994; Ptito, Brisson, Dell’Acqua, Lassonde, & Jolicœur, 2009). An important challenge is to unveil why some cognitive functions can be carried out independently in the separated hemispheres while other functions engage both hemispheres. Furthermore, it is now clear that accurate detection and localization is possible across the whole visual field, and there is some evidence that even more information concerning visual images can be transferred between hemispheres. Although we have some understanding of what types of information can be transferred in the visual domain, our knowledge base in the somatosensory domain is much more limited. This is probably due to a bias throughout cognitive neuroscience and psychology, leading to a strong focus on vision in split-brain research. It is important to collect converging evidence by investigating the somatosensory system which is also strongly lateralized. Note that in somatosensory processing transfer between hemispheres (about 80% correct for the bimanual conditions) has been observed for basic same-different matching of real objects (Fabri, Del Pesce et al., 2005).
Another important goal is to obtain a more detailed description of the perceptual, cognitive and linguistic capabilities of the disconnected right hemisphere. For understanding unity of mind, two capabilities specifically are crucial. First, experiments investigating aspects of the conscious mind often go beyond simple visual processing, and future studies will thus critically depend on testing high-level cognitive abilities of both hemispheres. Specifically, language abilities, crucial for understanding questions and instructions, will likely play a pivotal role. Thus, the first question is to what extent the right hemisphere is capable of language processing. Note that complicated instructions (Gazzaniga, Smylie, Baynes, Hirst, & McCleary, 1984; Pinto et al., 2017a; Zaidel, 1983), for instance relating to mental imagery (Johnson, Corballis, & Gazzaniga, 2001; Kosslyn, Holtzman, Farah, & Gazzaniga, 1985; Sergent & Corballis, 1990), seem to be well within the reach of the right hemisphere. Moreover, right hemisphere language capabilities seem to improve over time (Gazzaniga, Volpe, Smylie, Wilson, & LeDoux, 1979; Gazzaniga et al., 1996). Longitudinal language tests (for instance with a Token test: De Renzi & Vignolo, 1962) would further illuminate the extent of right hemisphere language processing.
Second, unveiling to what extent each hemisphere is capable of subserving consciousness at all seems relevant for unity of mind as well. If the disconnected right hemisphere can produce full-blown consciousness, then questions regarding unity of mind are clearly more pertinent then if the right hemisphere only produces minimal amounts of consciousness. Right hemisphere consciousness can be studied through novel neural paradigms (Bekinschtein et al., 2009; Casali et al., 2013; Pitts, Metzler, & Hillyard, 2014; Shafto & Pitts, 2015). For instance, Bekinschtein et al. employed EEG to measure if the brain detected irregularities (as indicated by an event-related potential [ERP] signal called the P3) in different states of consciousness. They found that when consciousness was reduced, local irregularities were still detected - for instance after three high auditory tones a low tone evoked a P3. However, global irregularities - several times a low tone followed three high tones, then on the critical trial three high tones were followed by another high tone - did not evoke a P3 when consciousness was reduced. Crucially, when consciousness was unimpaired both local and global irregularities evoked a P3 response. Right hemisphere consciousness may also be studied in other patient groups where interhemispheric communication is hampered. One particularly interesting group are post-hemispherotomy patients (Lew, 2014). These patients have been surgically treated to disconnect an entire hemisphere (usually for intractable epilepsy), but unlike hemispherectomy patients the disconnected hemisphere remains in place in the cranium and remains vascularized.
Clearly, the central question, whether each hemisphere supports an independent conscious agent, is not settled yet. Novel paradigms in this respect could lead to progress. For instance, a pivotal question is whether each hemisphere makes its own decisions independent of the other hemisphere. If each hemisphere produces its own autonomous conscious agent then this should be the case. That is, if two agents are asked to freely choose a random number, then the odds that they consistently pick the same number are small. And vice versa, if each hemisphere makes its own conscious decisions, independent of the other hemisphere, then this seems to rule out unity of mind. Note that each hemisphere making its own decisions is different from information processing occurring independently per hemisphere. Unconscious information processing is almost certainly split across hemispheres in a split-brain. However, this does not prove that consciousness is split or unified. Even in a healthy brain, where consciousness is unified, many unconscious processes run independently, and in parallel.
One way to tackle the central question is by having the hemispheres respond to questions in parallel. Overt behavior most likely does not allow for this, due to bilateral motor control processes sketched earlier. However, perhaps parallel responding is possible if the hemispheres produce covert responses. For instance, the patient could be asked to pick one of four options and indicate their choice by carrying out certain content-specific mental imagery tasks. This imagery can then be decoded in parallel from each hemisphere using neuroimaging techniques (see Owen et al., 2006 for a similar approach with vegetative state patients). If each hemisphere harbors an autonomous conscious agent, then it is highly unlikely that the two hemispheres will consistently make the same choices. Thus, if the choices are uncorrelated across hemispheres, then this may critically challenge the unified mind view.
Another way to tackle the question of unified consciousness in the split-brain is to employ ERPs as markers of concurrent conscious processing in the left and right hemispheres. For instance, in one study (Kutas, Hillyard, Volpe, & Gazzaniga, 1990) visual targets were presented either separately to the left or right visual field or to both visual fields simultaneously. It was found that the P300 - a signal possibly reflecting conscious processing of a visual target (Dehaene & Changeux, 2011; Dehaene, Charles, King, & Marti, 2014; Salti, Bar-Haim, & Lamy, 2012) - was reduced for bilateral targets. This suggests some type of integration of conscious processing. Studies employing ERPs may indicate whether conscious processing is unified, while unconscious processing is split, which would be suggestive of unified consciousness.
Posts
