Search Results

This chapter discusses a vision of computational neuroscience, and also addresses what has happened since the publication of David Marr’s Vision. It argues that it is now time to reemphasize the links between levels of understanding, if progress in computational neuroscience is wanted. It is believed that computational neuroscience over the past thirty years can be described as mostly exploring each level of understanding independent of the others. Some of the past research trends in computational neuroscience are explained. This brief look at computational neuroscience in the last decades indicates that significant progress has been made at each of the levels of understanding—in a sense following Marr’s prescription—though the problem is far from being solved. It argues that neuroscience can help computational theory and even computer science.

This book introduces a conceptual framework for brain function based on large populations of neurons. It advances the hypothesis that emergent properties are high-level effects that depend on lower-level phenomena in some systematic way, drawing on the idea that brains are computational in nature. Areas and topics related to computational neuroscience covered in this book include computational mechanisms in neurons, analysis of signal processing in neural circuits, representation of sensory information, systems models of sensorimotor integration, and computational approaches to plasticity. The book emphasizes the importance of single neuron models as the foundation into which network models must eventually fit. It also provides a background discussion on neuroscience and the science of computation.

This chapter provides an overview of the “neuroscience” component of the “computational neuroscience” synergy. It begins with a discussion of three ideas about levels in nervous systems: levels of analysis, levels of organization, and levels of processing. Levels of organization are essentially anatomical, and refer to a hierarchy of components and to structures that comprise these components. Levels of processing are physiological, and refer to the location of a process relative to the transducers and muscles. Levels of analysis are conceptual, and refer to different kinds of questions asked about how the brain performs a task. The chapter proceeds by considering seven categories of structural organization in nervous systems: systems, topographic maps, layers and columns, local networks, neurons, synapses, and molecules. It concludes by presenting a short list of brain facts.

Vast spectra of biological and psychological processes are potentially involved in the mechanisms of psychiatric illness. Computational neuroscience brings a diverse toolkit to bear on understanding these processes. This chapter begins by organizing the many ways in which computational neuroscience may provide insight to the mechanisms of psychiatric illness. It then contextualizes the quest for deep mechanistic understanding through the perspective that even partial or nonmechanistic understanding can be applied productively. Finally, it questions the standards by which these approaches should be evaluated. If computational psychiatry hopes to go beyond traditional psychiatry, it cannot be judged solely on the basis of how closely it reproduces the diagnoses and prognoses of traditional psychiatry, but must also be judged against more fundamental measures such as patient outcomes.

Advances in our understanding of brain function and dysfunction require the integration of heterogeneous sources of data across multiple levels of analysis, from biophysics to cognition and back. This chapter reviews the utility of computational neuroscience approaches across these levels and how they have advanced our understanding of multiple constructs relevant for mental illness, including working memory, reward-based decision making, model-free and model-based reinforcement learning, exploration versus exploitation, Pavlovian contributions to motivated behavior, inhibitory control, and social interactions. The computational framework formalizes these processes, providing quantitative and falsifiable predictions. It also affords a characterization of mental illnesses not in terms of overall deficit but rather in terms of aberrations in managing fundamental trade-offs inherent within healthy cognitive processing.

There is an ongoing philosophical and scientific debate concerning the nature of computational explanation in the neurosciences. Recently, some have cited modeling work involving so-called canonical neural computations—standard computational modules that apply the same fundamental operations across multiple brain areas—as evidence that computational neuroscientists sometimes employ a distinctive explanatory scheme from that of mechanistic explanation. Because these neural computations can rely on diverse circuits and mechanisms, modeling the underlying mechanisms is supposed to be of limited explanatory value. I argue that these conclusions about computational explanations in neuroscience are mistaken, and rest upon a number of confusions about the proper scope of mechanistic explanation and the relevance of multiple realizability considerations. Once these confusions are resolved, the mechanistic character of computational explanations can once again be appreciated.

John Moore initially became known for elucidating the action of tetrodotoxin and other neurotoxins using his innovative sucrose gap method for voltage clamping squid axon. He also was a pioneer in the nascent area of computational neuroscience, using computer simulations in parallel with experiments to predict experimental results and thus validate the concepts used in modeling. Intrigued by the possibility of applying his knowledge of physics to learn how neurons employ electricity to generate and transmit signals, he led the field in exploring how ion channels and neuronal morphology affect excitation and signal propagation. He developed electronic instrumentation of high precision for electrophysiology, the result of experience gained through an unconventional career path: early training in physics, assignments involving feedback in the Manhattan Project, and learning principles of operational amplifiers at the RCA Laboratories. His summers at the Marine Biological Laboratory in Woods Hole, MA, now exceeding 50, made much of his work possible and established the MBL as his intellectual home. In retirement, he developed the educational software Neurons In Action, coauthored with his wife Ann Stuart, that is now widely used as a learning tool in neurophysiology.

This concluding chapter sums up the key findings of this study on the computational neuroscience of vision. The results show that the responses of many inferior temporal visual cortex neurons have transform invariant responses to objects and faces, but not all neurons have view invariance. The findings also indicate that much of the information available from the responses of the neurons about shapes and objects is available in short time periods and that invariant representations can be self-organized using a trace learning rule incorporated in a feature hierarchy network such as VisNet.

Recurrent maladaptive patterns (RMPs) have been a foundational concept in psychodynamic therapy (PDT) and psychoanalysis for over a century. Typically associated with character pathology (i.e., personality disorders), they highlight the remarkable correspondences frequently observed between relationship patterns in a person’s family of origin, their current adult relationships, and the transference relationship with the therapist. These patterns can be understood as an expression of schemas and therefore share a common conceptual foundation with other major psychotherapy modalities. Yet, the centrality of affect in the origin/development of these maladaptive patterns and their treatment may not be widely appreciated among practitioners of PDT or any other modality. The basic thesis of this chapter is that RMPs as described in the PDT literature could potentially become more widely recognized, understood, and treated in an integrated manner if their developmental and affective origin were more generally appreciated. Doing so would not only improve interpersonal functioning but could also alter the affective dysfunction that predisposes to the development of symptoms that are a common reason for seeking treatment. Consistent with newer developments in psychodynamic theory grounded in observations from early childhood development, this chapter briefly reconsiders the fundamental elements of RMPs, including unconscious processes, development, conflict, defenses, and mechanisms of change from the perspective of affective science and computational neuroscience. In so doing, the goals are to broaden appreciation of the importance and ubiquity of RMPs by explaining them in nonclinical language, to increase the likelihood of enduring change by promoting an integrative approach to their treatment focusing on new emotional experiences in meaningful contexts and to facilitate research that can potentially establish the benefits of such an approach.

Clinical heterogeneity presents important challenges to optimizing psychiatric diagnoses and treatments. Patients clustered within current diagnostic schema vary widely on many features of their illness, including their responses to treatments. As outlined by the American Psychiatric Association Diagnostic and Statistical Manual (DSM), psychiatric diagnoses have been refined since DSM was introduced in 1952. These diagnoses serve as the targets for current treatments and supported the emergence of psychiatric genomics. However, the Research Domain Criteria highlight DSM’s shortcomings, including its limited ability to encompass dimensional features linking patients across diagnoses. This chapter considers elements of the dimensional and categorical features of psychiatric diagnoses, with a particular focus on schizophrenia. It highlights ways that computational neuroscience approaches have shed light on both dimensional and categorical features of the biology of schizophrenia. It also considers opportunities and challenges associated with attempts to reduce clinical heterogeneity through categorical and dimensional approaches to clustering patients. Finally, discussion will consider ways that one might work with both approaches in parallel or sequentially, as well as diagnostic schema that might integrate both perspectives.

The integrated memory model (IMM) proposed that the change process in psychotherapy involves the joint activation and reconsolidation of episodic memory, semantic memory, and emotional responses. The IMM did not thoroughly elaborate on what was meant by “emotional responses,” but a key concept was the distinction between implicit and explicit emotion. This chapter reviews the three-process model (TPM) of implicit and explicit emotion and its implications for extending the IMM. The TPM provides a detailed characterization (at cognitive, computational, and neural levels of description) of the processes associated with an emotional response. These processes include (a) situation appraisal and the subsequent generation of an affective (bodily, cognitive, and automatic skeletomotor) response, (b) the subsequent internal representation of that response (in terms of bodily sensations and emotion concepts), and (c) the role of salience, attention, and goal relevance in moderating whether or not one becomes aware of their emotions. After introducing the TPM, the author illustrates its utility in clarifying the nature of emotional responses in the IMM. The chapter also illustrates how the TPM can provide insight regarding the specific processes targeted by therapeutic interventions and how they could promote more adaptive emotional functioning.

A theory of cognitive rehabilitation should specify how change from a damaged state of cognitive processing can be modified into a normal, or more functional, state of cognitive processing. Such a theory should incorporate what is known about the cognitive representations and processes underlying normal cognition, how these are affected by brain damage, and how learning or modification of cognitive processing occurs. This chapter argues that development of a useful theory of cognitive rehabilitation requires integrating advances from cognitive neuropsychology, experimental psychology, computational neuroscience, and molecular biology of the brain, as well as empirical evidence from various branches of rehabilitation. It is likely that such a theory will specify how behavioral rehabilitation strategies can be augmented by pharmacological agents.

The introduction discusses the coverage of this book, which is about the computational neuroscience of vision. It introduces some of the background for understanding brain computation and discusses how some of the essential features of this can be captured by simple formalisms. The introduction also explores three neuronal network architectures, long-term potentiation and long-term depression, and the fine structure of the cerebral neocortex.

This chapter discusses psychoneural identity theory and how it aims to discredit sensory phenomenology—or qualia—for philosophical realist and reductionist programs. Paul and Patricia Churchland’s works are exemplary of such motivation. Paul Churchland’s philosophizing of computational neuroscience attempts to resolve mental contents into vector coding and its transformations, yet what he describes is not phenomenology but a sensory schema of psychology. Patricia Churchland, on the other hand, admits there are few or no intertheoretic identities, and therefore no proper analogies from them to her projected psychology-to-neuroscience reduction. Their misrepresentations of the nature of perception are documented here, with a conclusion that the dogmatic denial of phenomenology by reductionist philosophy heretofore involves invalid and unsound arguments.

Sense of agency—the feeling of being the author of one’s actions—may be a critical component of one’s sense of self and of one’s interaction with the world. Insights from clinical and experimental neuropsychology, as well as cognitive and computational neuroscience, have provided complementary evidence that the sense of agency arises from the integration of an array of internal and external cues. These frameworks can help to explain how disruptions in one or more of these cues may result in altered experiences of agency. This chapter reviews these explanatory frameworks and shows how important and useful they have become in making sense of an array of clinical observations, from the disorders of control and agency that result from circumscribed brain damage to the widespread attenuation of agency that may characterize psychosis in which no clear brain lesion has been identified.

This chapter provides an overview of biological perceptual systems and their underlying computational principles focusing on the sensory sheets of the retina and cochlea and exploring how complex feature detection emerges by combining simple feature detectors in a hierarchical fashion. We also explore how the microcircuits of the neocortex implement such schemes pointing out similarities to progress in the field of machine vision driven deep learning algorithms. We see signs that engineered systems are catching up with the brain. For example, vision-based pedestrian detection systems are now accurate enough to be installed as safety devices in (for now) human-driven vehicles and the speech recognition systems embedded in smartphones have become increasingly impressive. While not being entirely biologically based, we note that computational neuroscience, as described in this chapter, makes up a considerable portion of such systems’ intellectual pedigree.