Connectivity and cognition: the function of white matter

The role of white matter in human behavior is surely a vast and imposing topic. While many in neuroscience and cognitive neurology are deterred from the study of white matter because its role is considered merely supportive and the cerebral cortex mediates the higher functions, such a view is incomplete. The simple fact that about half the human brain is made up of white matter immediately compels respect for what has long been overlooked in the study of cognition and emotion (Filley, 2010). Moreover, as discussed in Chapter 3, white matter has expanded more in evolution than cortical gray matter (Zhang and Sejnowski, 2000), an observation that has helped foster the current view that the human brain owes its singular capacities not only to its large number of neurons (Rushton and Ankney, 2009) but also to the connectivity conferred by myelination (Roth and Dicke, 2005). The impressive evolutionary expansion of myelinated tracts, especially those intimately involved with the frontal lobes (Zhang and Sejnowski, 2000; Schoenemann, Sheehan, and Glotzer, 2005; Smaers et al., 2010), offers an important clue that the behavioral neurosciences will discover only partial understanding by focusing on the cerebral cortex alone.

Clinically, beyond the useful generalization that white matter facilitates the operations of distributed neural networks, in most cases the specific neurobehavioral details of tract function are yet to be worked out. Excepting such large tracts as the corpus callosum, neuroradiologists in clinical practice are usually unable to identify specific white matter connections so that the neurobehavioral effects of lesions can be observed. Whereas myelination clearly increases conduction velocity and improves network efficiency, white matter tracts will be increasingly found to carry out more specific functions in the panoply of cognitive and emotional operations, contributing to behavior by virtue of their location within networks dedicated to cognitive and emotional functions. White matter participates as an essential neuroanatomic component of cortical and subcortical structures linked in reciprocal relationships enabling the phenomena of cognition and emotion (Parvizi, 2009). Thus a myelinated connection will be uniquely contributory not because of its myelination alone but also because of its position within a network devoted to a given operation such as memory, language, or executive function. In this regard, white matter adheres perfectly to the neurobehavioral dictum that the location of the structure predicts its function. Moreover, by implication, the location of neuropathology determines the nature of the clinical deficit. Elucidation of the details of white matter connectivity will thus steadily expand the understanding of how the brain mediates singularly human capacities and their dissolution in neuropathological states (Filley, 2012).

The study of dementia in the context of white matter can be seen as one aspect of the broader issue of how white matter contributes to the higher functions of the human brain. WMD is a concept intended primarily to underscore the crucial part played by myelinated systems in higher function, not to instantiate an immutable diagnostic category that will be clearly distinguishable by clinical evaluation. If the idea of WMD, together with the related construct of MCD, encourages investigation of the means by which the brain exploits its exquisite connectivity to facilitate the myriad operations of human behavior, its purpose will have been well served.

A useful summary statement is that white matter subserves the essential function of information transfer in the brain as a complement to the information processing of gray matter. No doubt exists that the elaborate mechanisms of synaptic function within gray matter enable the extraordinary processes that have been linked with higher functions such as memory acquisition, and conceptualized in terms such as long-term potentiation. Equally vital are the pathways that link gray matter regions into coherent functional networks that permit the rapid and efficient activity of integrated cognitive and emotional operations (Mesulam, 1990; Catani, 2006; Wang et al., 2015).

The study of white matter and cognition also informs the long and only tenuously resolved debate in cognitive neuroscience regarding holism versus localizationism. If white matter is important in the operations of human behavior, does it function as an equipotential whole, or are individual tracts associated with specific functions? This question was raised in past eras with respect to the cerebral cortex in the controversy about whether the cortex mediates cognition in an undifferentiated manner, as Karl Lashley and Henry Head would contend, or contains specific regions invariantly dedicated to specific functions, as argued by Franz Joseph Gall and later Paul Broca. The debate has been tentatively settled by the recognition that while individual cortical regions are indeed seen as important, and focal syndromes do occur with focal damage, the entire cortex shares a certain commonality of function, and diffuse degeneration such as that seen with AD leads to widespread cognitive dysfunction. In the case of white matter, it is likely that the same solution to this debate will obtain; that is, whereas the entirety of white matter can be seen as important in the overall syndrome of WMD, certain regions can be seen as important for specific functions (Filley, 2011).

The key concept that helps resolve both debates is the idea of distributed neural networks, the collections of gray and white matter structures devoted to individual neurobehavioral functions (Mesulam, 1990). Disruption of these networks, whether involving gray or white matter, leads either to an isolated syndrome such as aphasia or executive function, or to a more pervasive syndrome of impairment such as AD or WMD. The challenges of understanding these networks are not to be met by stale arguments about holism versus localizationism, but by acknowledging that both focal and diffuse syndromes can be produced by brain disease, and that the contributions of both gray and white matter involvement must be included for a complete account.

Disconnection and leukocentrism

In 1965, Norman Geschwind famously championed the notion of disconnection, and made it clear that white matter deserves attention in the neuroscientific study of human behavior (Geschwind, 1965). His paper can thus be seen as inaugurating an era of leukocentrism. The years following this work have gradually but unequivocally demonstrated the prescience of his ideas. Geschwind was intrigued with specific connections made within the brain, and the neurobehavioral effects of disconnecting lesions. Realizing that gray matter regions as well as white matter tracts could act as relay stations within cerebral networks, he chose to focus more on cortical structures and less on white matter. Indeed, the term “white matter” appears only infrequently in his 1965 paper, and not a single diagram or illustration of brain connectivity appears in this work of well over 100 pages. But white matter tracts were clearly a part of his thinking, even if the details of their anatomy and pathology were often obscure. The cognitive effects of selective tract disruption were included in his views of the brain and behavior, as they had been in the previous century by many European neurologists.

After Sigmund Freud was largely responsible for turning neuroscientific attention away from the localization of behavior in the early twentieth century, Geschwind restored interest in brain–behavior relationships with a spirited reconsideration of the brain and its connectivity. He viewed this focus as critical to the understanding of behavior, since studying a patient without knowing the state of the brain, as Freud had made the standard approach for many decades of the twentieth century, would inevitably be insufficient. As Geschwind wrote a decade after the appearance of his disconnection paper, “It must be realized that every behavior has an anatomy” (Geschwind, 1975).

As more information has been gathered in the age of neuroimaging, however, the simultaneous involvement of many tracts has been found to be much more common than isolated tract disruption. Indeed, with some notable exceptions, it is rare in clinical neurology to encounter a single, well-defined white matter lesion that can be usefully studied with regard to its cognitive effects. Instead, clinicians are regularly confronted by patients in whom many lesions are simultaneously apparent. The consequences of such multifocal involvement immediately complicate the analysis of clinical effects, but a concerted effort from the neuroscientific community to take on this task cannot be avoided. Just as the effects of cortical disease, broadly envisioned, are routinely considered in the approach to dementia, so too should white matter disease be included in the discussion. In short, the idea of leukocentrism – which was at first founded on the study of isolated tracts, their lesions, and their clinical consequences – now implicates the entirety of white matter in the brain. One of the ways this diffuse form of white matter involvement is being studied is with respect to the major psychiatric diseases (Walterfang et al., 2005; Haroutunian et al., 2014), itself a worthy endeavor that may reveal major insights into mental illness. The task of this book, however, has been to show how this line of investigation unavoidably leads directly to the study of dementia.

The clinical and research benefits of investigating WMD may be substantial, and far more apparent than can be demonstrated at present. The diagnosis of many patients with white matter disorders affecting cognition can be enhanced, particularly if the precursor syndrome of MCD proves useful in identifying those with early involvement. Treatment, involving existing and many evolving modalities, will continue to advance, and the opportunity to treat at the early stage of MCD may substantially improve outcome. Research will be invigorated by a host of innovations that consider the white matter as the primary site of neuropathology, and as a tissue in which specifically targeted therapeutic interventions may be transformative. One of the key areas that will assume increasing importance is the role of intracortical white matter, as these small fascicles have not received the attention devoted to the larger tracts coursing within and between the hemispheres and to and from more caudal brain regions. Many insights are likely from the investigation of this largely unexplored field.

Study of the concepts of WMD and its companion MCD may also reveal new ways of conceptualizing major neurobehavioral disorders that are either idiopathic or poorly understood. AD is the most obvious example of how a leukocentric perspective may radically restructure the field and inform new efforts to detect the initial neuropathological insult and devise new and more effective treatments. Meanwhile, CTE is an example of a disease in which the initial insult is known but the details of pathogenesis later in the course are not clear. In both diseases, the role of protein misfolding and interneuronal propagation, intensely investigated as phenomena with potentially widespread relevance to a host of neurodegenerative processes (Prusiner, 2012), may in fact become clarified as a later effect of some primary process determined largely by acquired factors well before the onset of clinical symptoms. In contrast to what is thought typically to be a stochastic process, by which proteopathies develop as a random occurrence (Prusiner, 2012), the spread of deleterious proteins to various regions around the brain may often be found to relate to common and potentially reversible factors mediated through white matter pathology such as hypertension and traumatic brain injury. Meanwhile, from the world of neuropsychiatry, schizophrenia and autism stand out as examples of highly prevalent and disabling disorders in which an interpretation involving the fundamental disturbance as a connectopathy may be illuminating.

In short, leukocentrism offers the exhilarating prospect of a new way of thinking about the brain and behavior. What is the logic of ignoring one half of the brain in thinking about cognition? Is the greater enlargement of white matter compared to gray in evolution to be disregarded? Surely a cardiology that considers only the left side of the heart to be important, for example, would not long prove productive. There is much to be learned by investigation based on curiosity about how a leukocentric approach can be instructive. To invoke one example, it has recently been recognized that plasticity in the white matter may be an important foundation of learning and memory (Fields, 2010).

Still more intriguingly, the phenomenon of creativity, an area of cognitive performance widely cherished but imposingly difficult to study, has been investigated using DTI, and significant relationships have been disclosed between creativity as measured by a divergent thinking test and fractional anisotropy in a number of association tracts and the corpus callosum (Takeuchi et al., 2010). Given the widely held view that creativity generally involves the conjoining of seemingly unrelated concepts into a novel synthesis (Austin, 1978; Heilman, 2005), it would follow that studying brain connectivity via white matter tracts is a promising approach to understanding this cognitive capacity. A related and fascinating observation is that Albert Einstein’s brain features a corpus callosum that is notably thicker than those of age-matched control brains, particularly in posterior callosal regions connecting the parietal lobes (Men et al., 2014), which have also been found to be exceptionally well developed in this creative genius (Witelson, Kigar, and Harvey, 1999). Could it be that the universally exalted capacity of creativity will be found to depend critically on the integrity of tracts interconnecting gray matter regions and enabling the formation of novel insights by enhancing the physical integration of disparate neural ensembles?

The notion that white matter tracts operate as integrative structures allowing for cooperation between gray matter regions is appealing indeed, and may inform the study of many questions in cognitive neuroscience that are not readily answered by a reductionistic focus on one or another cortical zone. Corticocentrism should of course not be put aside, consigning gray matter to the same fate that white matter has endured. But it should be combined with its leukocentric counterpart, allowing the fully integrated study of all the brain in the performance of its impressive operations.

Reflections on the study of white matter and cognition

In the course of examining the contributions of white matter to cognitive function, it becomes abundantly clear that this project involves not just clinical but also basic neuroscience. While the concepts of WMD and MCD surely provoke a broad range of clinical implications – prevention, diagnosis, prognosis, and treatment – they also require an understanding of the fundamental aspects of white matter as a component of the brain. These approaches, which can be considered top-down and bottom-up respectively, serve best when complementing each other, and the most thorough understanding derives from considering the tissue in which neuropathology arises as well as the clinical phenomenology produced.

Extending beyond the corticocentric perspective of much contemporary neuroscience, the study of white matter and cognition offers the invigorating prospect of uniquely furthering the pursuit of brain–behavior relationships. As described in Chapter 5, the process of integrative review (Grimes and Schulz, 2002; Whittemore et al., 2014) based mainly on the lesion method of behavioral neurology (Damasio, 1984) offers a legitimate means of exploring how a major component of the brain contributes to the operations of human behavior. As a medical concept, WMD qualifies as a topic of applied science (Davis, 2000), as it is anticipated that understanding the clinical syndrome will lead to interventions that can alter the outcome of human beings at risk for or enduring the clinical effects of white matter disorders. The practical applications of this perspective may have far-reaching implications in terms of medical care and public policy intended to reduce the human suffering produced by white matter disorders.

Importantly, however, the study of WMD can also be used to infer the structure and function of normal white matter, thus serving the essential goal of basic science, which is to understand the natural world (Davis, 2000). Indeed, the distinction between basic and applied science is often not clear (Davis, 2000), and in the study of dementia as well as other related syndromes, behavioral neurology has in fact flourished at the interface of these endeavors. The clinical task of caring for people who are facing the cognitive consequences of brain disorders necessarily involves the integration of basic neuroscience, and the two approaches complement and reinforce each other. The medical imperative of helping people cope with potentially devastating conditions has comfortably coexisted with the scientific goal of understanding the brain through detailed examination of the effects of its disorders on normal structure and function.

The relationship of the WMD concept to basic neuroscience has a special distinction in that the vast complexity of human cognition must be considered. It is not possible to fully understand the complexities of human behavior by only using experimental paradigms specific to isolated mental operations, and understanding the entire organism must ultimately be incorporated in a complete account. As a critical component of the brain that directs the organism’s behavior, white matter as a whole demands focused study. As imposing as it may seem, this task is unavoidable.

Basic science necessarily involves reductionistic investigation in the laboratory, where a tightly defined problem can be isolated, described, and modeled as a basis for understanding a larger system. It is of course indisputable that enormous advances have been made in science by the use of reductionism, and many more are sure to follow. To call upon an example from the neurosciences, the histologic study of the brain, with its identification of abnormal proteins and the like, does indeed serve to inform the study of brain–behavior relationships (Mesulam, 2012) in concert with sophisticated neuroanatomic (Schmahmann and Pandya, 2006) and neuroimaging (Catani and Thiebaut de Schotten, 2012) studies that contribute to mapping brain connectivity. Similarly, it is crucial to appreciate from a neuropathological perspective, for example, that axonal damage in white matter substantially worsens the prognosis compared to lesions in which only myelin is damaged (Trapp et al., 1998; Medana and Esiri, 2003).

However, when the primary objective is the study of the impressive spectrum of cognitive operations, human beings in their extraordinary complexity must be the central focus of investigation (Damasio, 1984; Bear, 1997). This approach falls squarely within the realm of systems biology (Villoslada et al., 2009), an emerging field that adopts an integrative strategy to understand higher-level operating principles of living organisms (Villoslada et al., 2009). Systems biology, based on the notion of biological networks in which emergent properties can be derived from a consideration of both structure and dynamics (Villoslada et al., 2009), is ideally suited to inform the investigation of white matter as it functions to subserve human behavior in health and is altered in disease (Haroutunian et al., 2014).

Experimental laboratory studies focused on specific molecular, genetic, or pharmacological aspects of human behavior are not enough. Indeed, the classic reductionistic approach to biomedical investigation, while doubtless adding important information, has not produced comprehensive understanding of the pathogenesis and effective treatment of highly prevalent and threatening disorders such as AD (Villoslada et al., 2009). Many clinicians also tout the value of the meta-analysis, in which numerous quantitative investigations are rigorously combined by sophisticated statistical manipulation, but this too falls short in some respects.

Rather, it is critical to understand the cerebral origins of behavioral dysfunction by assembling the sequelae of structural damage from many common and uncommon disorders, and then seeking to combine these data into a coherent clinical syndrome with wide generalizability. Cataloguing and synthesizing the cognitive effects of white matter dysfunction can thus be seen as basic science as much as examining the impact of a gene mutation, a viral infection, or an inflammatory cascade. The reductionistic approach that has so dominated recent neurobiological investigation must be complemented by a systems approach founded on the principle that the whole is more than the sum of its parts (Villoslada et al., 2009).

As the two parallel inquiries – top-down and bottom-up – proceed in what will ideally be a mutually useful collaboration, the distinction between basic and clinical research in neuroscience becomes indistinct and increasingly irrelevant. To illustrate, the fact that the amyloid plaque is a feature of AD does not mean that it is an explanation, and the elucidation of the etiopathogenesis of AD may well demand a broader systems biology approach that considers the neural networks in which the disease first begins. The same reasoning can be applied to the problem of CTE, in which the neuropathological observation of cortical tau deposition may be a function of events affecting brain connectivity many years before symptoms even appear. As much as can be learned about one or another isolated protein, a reductionistic focus on these molecules may fail to identify the larger picture needed to understand a difficult problem. This broadly synthetic way of thinking has inspired and sustained the integrative review over many years underlying this monograph and its closely related predecessor (Filley, 2012).

Much remains to be accomplished, and many long-cherished ideas about the role of the cerebral cortex in higher functions will require a fresh look and critical reevaluation. The notion of the connectome not only highlights neuroscientific efforts in this quest over many centuries but also helps organize systematic study of the structure and function of normal and abnormal white matter with respect to cognition (Geschwind, 1965; Sporns, 2011; Catani et al., 2013). Considering the contributions of white matter to cognition and its decline significantly expands the study of human mentation that is the essence of behavioral neurology.