Psychiatric Neuroscience: Incorporating Pathophysiology into Clinical Case Formulation

Brain Lesions and Behavior

There is a strong tradition within classical neuropsychology and behavioral neurology of understanding neuroanatomical circuitry by studying the emergent or lost behaviors in patients with focal brain lesions. These studies have provided us with a rich view of various brain regions and their relationship to behavior. Perhaps the most famous case is that of Phineas Gage, the Vermont railway worker who suffered a traumatic lesion bilaterally to the medial frontal lobes and developed personality changes. Another famous patient (known by his initials) is H. M., who underwent bilateral medial temporal lobe resection for intractable epilepsy and as a result lost the ability to form new declarative memories. While striking and informative, findings from these rare cases may be difficult to extrapolate to the pathophysiology of common psychiatric illnesses, which generally do not involve focal lesions. Traditionally, biological psychiatry has relied more on biometrics and quantitative methods; these population-based approaches risk losing insights available from rare cases but are more likely to produce broadly generalizable findings.

Neuropsychology and Endophenotypes

An increasingly important approach in psychiatric neuroscience is to identify and study intermediate phenotypes. These are quantitative phenotypes that are closely associated with the clinical syndrome of interest, but which are less complex and easier to link to the function of specific neural circuits. They can also be used to identify biologically relevant subtypes within a diagnostic category, reducing heterogeneity that may limit the power of scientific investigations. Endophenotypes are intermediate phenotypes that are present both in affected individuals and in their unaffected relatives, therefore reflecting genetic risk independent of actual disease. Neuropsychological tests of cognitive function are commonly used to identify endophenotypes. For example, impairment of working memory, which is closely related to the function of dorsolateral prefrontal cortex, is found within a subgroup of patients with schizophrenia. Endophenotypes thus help bridge the gap between brain circuits, which are amenable to study at the molecular and cellular level, and clinical syndromes, which are less tractable. This approach becomes especially powerful when combined with other methods, such as neuroimaging or genetics.

Neuroimaging

Neuroimaging has provided one of the best modern tools for examining the pathophysiology of mental illness in the living brain. Neuroimaging can provide many different quantitative measures (including morphometry, metabolism, and functional activity). Neuroimaging research using groups of subjects can determine whether mental illness is associated with changes in the size or shape of specific brain regions, the functional activity within these regions, or their concentration of particular neurotransmitters, receptors, or key metabolites. Although neuroimaging methods can be used to measure cellular and molecular phenomena, the currently achievable spatial resolution still represents an important limitation in examining the microscopic pathological changes implicated in psychiatric illness.

Neurophysiology

There is a strong tradition within psychiatric neuroscience of studying the electrical activity of the brain and its relation to function. These methods include electroencephalography (EEG), event-related potentials (ERPs), and, most recently, magnetoencephalography (MEG), and transcranial magnetic stimulation (TMS). Like functional neuroimaging, these modalities provide information about the living, functioning brain. At present, electrophysiological techniques cannot provide anatomical resolution at the level of neurochemistry or synaptic physiology, and are limited to the study of cortical phenomena; however, they can provide excellent temporal and spatial resolution and are invaluable in studying the coordinated function of widely distributed neural circuits. Abnormalities in the timing of oscillations in neural circuit activity have been associated with psychiatric illnesses, and this is an area of intense research activity. For example, the reduction of gamma frequency (30 to 80 Hz) oscillations in schizophrenia has been ascribed to impaired N- methyl- d -aspartate (NMDA) receptor activity on GABA-ergic interneurons. These non-invasive methods are particularly heavily used in studies of brain development and function in children.

Brain Stimulation and Neuromodulation

Brain stimulation and neuromodulation techniques encompass a variety of device-based methodologies able to generate focal electrical currents in pre-selected brain regions. These currents are able to increase or decrease the excitability of the target neurons, and modulate the networks they belong to by acting as a neural pacemaker.

Brain stimulation can be divided among invasive and non-invasive approaches. Invasive techniques require the surgical implantation of stimulating electrodes in the brain, and are therefore exclusively used in therapeutic settings where the risk/benefit analysis is favorable. They include deep brain stimulation (DBS) and vagus nerve stimulation (VNS). Non-invasive methods do not require surgery or anesthesia, are very safe, and alter brain function in ways that are transient and reversible. The better-known and most commonly used methods are transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS). Chapter 18 describes these methods and their therapeutic applications in detail.

TMS has been used since the mid 1980s as a tool to study brain structure and function. Event-related paradigms using single pulses time-locked to a given stimulus or task have been used to determine the chronometry of the computations in a given brain region with great temporal resolution (in the order of milliseconds). Repetitive TMS (rTMS) can increase or decrease the excitability of a given area beyond the time of stimulation, creating a “virtual lesion” that lasts 15–60 minutes after the stimulation. This virtual lesion approach has been used, following the tradition of classical lesion studies, to understand the functional role of discrete brain regions.

Although neuroimaging and electrophysiological techniques are defined by their spatial and temporal resolution, what sets brain stimulation methods apart is their causal resolution . Neuroimaging and electrophysiological methods are observational; they measure patterns of brain activity (the dependent variable) in the context of a given task or disease state (independent variable). Such a design is able to establish correlations among these measures, but can never determine that a given pattern of brain activity is causing a mental state (or vice versa). On the other hand, brain stimulation techniques are interventional. They modify the system by changing brain activity (now the independent variable) and measure the behavioral, cognitive or affective changes that follow. This design offers causal explanatory power, which makes it a useful tool to answer a number of questions.

Neuropathology

Many researchers examine post-mortem neural tissue from those who suffered from psychiatric illness during their life-time. Post-mortem analysis reaches a level of molecular and cellular resolution currently unachievable in vivo ; however, it is commonly limited by confounds (such as age, effects of chronic medication, and non-specific effects of chronic psychiatric illness).

Neuropathology was clearly in fashion in the late 1800s and early 1900s, when Alzheimer first described plaques in the brain of his patient with dementia, and identified frontal cortex abnormalities in schizophrenia. While some skeptics have described schizophrenia as the “graveyard of neuropathologists,” recent studies have actually provided reproducible descriptions of deficits (such as those in parvalbumin-expressing GABA-ergic interneurons in deep layers 3 and 4, akin to Alzheimer's findings) in the cortex. These neuropathological findings have provided one of the strongest etiological hypotheses for schizophrenia.

Psychopharmacology

More than any other methodology in psychiatric neuroscience, pharmacology has been used to understand the neurochemical basis of behavior and to develop hypotheses regarding psychopathological mechanisms. Famous examples include the dopamine and glutamate hypotheses of schizophrenia, the catecholamine depletion hypothesis of depression, and the dopaminergic models of attention-deficit/hyperactivity disorder (ADHD) and substance abuse. In relating pharmacological effects to potential disease mechanisms, it is important to note that the effects of drugs on clinical symptoms may reflect mechanisms that are downstream of the core pathophysiology, or even unrelated to core disease mechanisms. By analogy, diuretics can improve the symptoms of congestive heart failure while providing less direct insight into its core pathophysiology. Nonetheless, by clearly connecting cellular and synaptic mechanisms with clinical symptoms, pharmacology provides mechanistic tools and information with enormous clinical and scientific utility.

Animal Experiments

In the authors' opinion, the value of animal experiments has received too little emphasis in psychiatric neuroscience. Clearly, complex psychiatric symptoms (such as delusions) cannot be modeled well in animals, and anthropomorphic interpretations of animal behavior should be taken with due skepticism. Despite these caveats, animal behaviors with known neuroanatomical correlates have been critical in elucidating the neurocircuitry and neurochemistry underlying many psychiatric phenomena. For example, anxiety- and fear-related behaviors have been very productively modeled in animals, leading to a detailed understanding of the role of the amygdala in these behaviors. Of course, animal studies also permit a wider range of experimental perturbations than possible with human investigations. Independent of their value as behavioral models, animal models therefore offer the opportunity to explore cellular and molecular pathophysiology in ways that are ethically or technically impossible in human subjects. For example, the fragile X knock-out mouse is an excellent model for fragile X syndrome, the most common form of inherited cognitive impairment. Studying these mice has led to a deep understanding of relevant defects in dendrite formation and neurophysiology.

Human Genetics and Molecular Biology

Adoption, twin, and familial segregation studies have proven that many psychiatric conditions are highly heritable (i.e., caused in large part by the additive effect of genes). Genetic endeavors in psychiatric neuroscience may be broken up into two broad categories: “forward genetics,” or genome-wide attempts to identify genetic loci (genes or their regulatory elements) that underlie susceptibility or contribute to pathophysiology; and “genotype-phenotype” studies, whereby candidate genes are chosen based on a priori biological hypotheses and the degree to which a gene plays a role in a given phenotype is assessed. Such phenotypes may be clinical diagnoses, or endophenotypes from neuropsychology or neuroimaging. The promise for human genetics in psychiatry is tremendous, especially for forward genetics, wherein researchers may be led to the core pathophysiology without requiring any a priori hypotheses. Yet human genetics research is exceedingly challenging for various reasons that are beyond the scope of the current discussion. In brief, the genetic architecture of neuropsychiatric conditions is heterogeneous and complex. That is, the majority of psychiatric illnesses likely reflect complex interactions of multiple genes, as well as their interaction with environmental factors that are difficult to assess. Despite these difficulties, there have already been a few notable examples of success.

Analysis of rare, large families with early-onset dementia led to the discovery of mutations in amyloid precursor protein (APP) and presenilins in Alzheimer's disease (AD). In these rare families, these mutations are statistically “linked” to disease and considered “highly penetrant.” However, the vast majority of AD patients do not have mutations in these genes. Indeed, in psychiatry examples of highly penetrant, simple dominant or recessive gene mutations are rare. That is, examples in neuropsychiatry of a particular gene mutation “causing” a specific condition are exceedingly rare and somewhat controversial, and the generalizability of these findings to the common conditions with more complex inheritance is usually unclear. Nonetheless, the APP pathway has provided an important target for drug development, which may lead to a medicine that stalls the progress of disease.

Genetic association studies through population genetics represent another approach to identifying susceptibility genes in “forward genetics.” In this approach, a common variation in the genome, such as single nucleotide polymorphisms (SNPs), is assessed for a statistically significant association with illness. This approach is based on the so-called common disease–common variant hypothesis. That is, psychiatric disorders may be in part due to a disadvantageous combination of a number of common forms of genetic variation as opposed to frank deleterious mutations. Again, a successful example of a gene association in neuropsychiatry comes from the field of AD wherein there is a fairly robust association at the level of population genetics or epidemiology between a common polymorphism in ApoE4 and susceptibility to AD. However, sometimes even when genetic associations are robust, the amount of the phenotype (i.e., the variance) that is explained by the given gene may be small and therefore the role in causation may be indirect or unclear. In addition, association studies have often used small sample sizes and thereby risked false-positive findings or problems of reproducibility. Now, appropriately-powered association studies (involving thousands of subjects) are underway; many of these use new and more powerful high-density genotype methods, namely “SNP chips” or microarrays. Even still, critical insights into the causative roles of genes in some psychiatric illness may only come from studies of gene–gene or gene–environmental interaction, or by using endophenotypes (such as neuroimaging) that may be closer to the action of the gene.

Methodologies in molecular genetics and molecular neuroscience also promise improved understanding of gene function in the brain. These methods include the following: comparison of gene sequences in human to non-human primate and other animals ; a deeper understanding of how non-coding elements within the genome may regulate important brain genes and thereby play a role in psychiatry; the study of gene expression using microarrays ; the study of gene function in mice in which specific genes have been modified by recombinant methods (e.g., “knock-out” or “knock-in” studies) ; and studies examining how experience and the environment alter gene expression. In summary, genomics and molecular genetics hold great promise for identifying genes and thus biological mechanisms at the core of psychiatric pathophysiology.

Process

A key concept in basic neuroscience and its clinical specialties is neuroplasticity. Although it is defined in different ways and can be studied at various levels of resolution (e.g., circuits, synapses), this term generally refers to the capacity of the neural system to change in response to external or internal stimuli following predetermined rules. Neuroplasticity provides a great deal of flexibility and adaptive capacity to the brain, permitting variable computational strategies and patterns of connectivity in a changing environment. Despite the significant potential for reactive (and adaptive) change, this happens around an exquisitely regulated homeostatic equilibrium point. Nevertheless, when the plastic changes are restricted, excessive, or occur around an altered equilibrium state, pathology develops. Luckily, the brain remains plastic, and any intervention (e.g., medications, psychotherapy or brain stimulation) that is effective in changing pathological cognition, behavior or affect induces adaptive plasticity. That is, a pathological mental state is sustained by a given pattern of brain activity, and changing this mental state will require changing its associated neural computational algorithm. Therefore, neuroplasticity is a key dynamic property of the brain that allows adaptive change (including learning and memory), but it is also an important source of pathology, and a necessary mechanism of action of effective neuropsychiatric treatments.

Although the specific pathophysiological mechanisms that lead to neuropsychiatric disease are many, we will consider two relevant examples: neurodevelopment, and neurodegeneration. Under neurodevelopment we include related processes that continue into adulthood (such as neurogenesis). Previously underestimated, adult neurogenesis is now known to continue in select regions of the human brain, most notably the olfactory bulb and the hippocampus. Although the role of adult neurogenesis in humans remains largely unknown, some evidence has connected altered hippocampal neurogenesis to mood disorders.

Neurodevelopmental processes shaping brain circuits have life-long effects on patterns of affect, behavior, and cognition with direct relevance to mental health. The effects of childhood experience have always been central to psychiatric understanding; psychiatric neuroscience has also attempted to provide a biological grounding for this understanding. Thus, neurodevelopmental processes include the interacting effects of genes and environment on brain and behavior. Figure 1-1 shows the processes of brain development, including intrauterine neuronal patterning, neurogenesis, cortical migration, gliogenesis, myelination, and experience-dependent synapse modification. In the first years and decade of life, the brain undergoes a process of synapse formation and pruning. Initially, neurons form an over-abundance of synapses that are then strengthened and pruned possibly based on experience, learning, or aging ( Figure 1-2 ).

Specific psychiatric disorders may be framed in terms of one or more of these three mechanisms. Autism or attention deficit hyperactivity disorder are examples in which a process of brain development goes awry. At the other end of life, neurodegenerative processes dominate, and can lead to dementias (e.g., Alzheimer's or frontotemporal lobar degeneration) or movement disorders (such as Parkinson's disease). Substance use disorders may reflect a combination of both processes modulated by maladaptive plasticity. Patients with substance dependence may have a susceptibility based on neurodevelopment, including a predisposition to risk-taking behaviors. Substance abuse also causes neuroplastic changes at the level of the synapse. Finally, chronic use of substances can cause neurodegeneration and dementia.