It has been recognized for a long time that concussions are dangerous injuries with potentially life-changing consequences, ranging from permanent symptoms to degenerative neurological states. The intellectual history of medicine and science from 1870 to the recent past shows both a continuity of clinical observations about HI and a steady, incremental accumulation of knowledge refining our understanding of those observations from a remarkably wide sphere of scientific disciplines.

No abstract

The Minnesota Multiphasic Personality Inventory (MMPI) was developed at the University of Minnesota, Minneapolis, in the 1930s and 1940s. It became a highly successful and highly controversial psychometric tool. In professional terms, psychometric tools such as the MMPI transformed psychology and psychiatry. Psychometric instruments thus readily fit into the developmental history of psychology, psychiatry, and neurology; they were a significant part of the narrative of those fields' advances in understanding, intervening, and treating people with mental illnesses. At the same time, the advent of such tools also fits into a history of those disciplines that records the rise of obsessional observational and evaluative techniques and technologies in order to facilitate patterns of social control that became typical during the Progressive Era in the United States and after. It was those patterns that also nurtured the resistance to psychometrics that emerged during the Vietnam War and after.

The learned weights of a neural network have often been considered devoid of scrutable internal structure. In this paper, however, we look for structure in the form of clusterability: how well a network can be divided into groups of neurons with strong internal connectivity but weak external connectivity. We find that a trained neural network is typically more clusterable than randomly initialized networks, and often clusterable relative to random networks with the same distribution of weights. We also exhibit novel methods to promote clusterability in neural network training, and find that in multi-layer perceptrons they lead to more clusterable networks with little reduction in accuracy. Understanding and controlling the clusterability of neural networks will hopefully render their inner workings more interpretable to engineers by facilitating partitioning into meaningful clusters.

The attitudes that characterize the contemporary "neuro-turn" were strikingly commonplace as part of the self-fashioning of social identity in the biographies and personal papers of past neurologists and neuroscientists. Indeed, one fundamental connection between nineteenth- and twentieth-century neurology and contemporary neuroscience appears to be the value that workers in both domains attach to the idea of integration, a vision of neural science and medicine that connected reductionist science to broader inquiries about the mind, brain, and human nature and in so doing supposedly resolved once and for all questions germane to the human sciences, humanities, and arts. How those attitudes were produced and reproduced first in neurology and then in neuroscience; in what way they were constructed and disciplined, thereby eventuating in the contested sciences and medicines of the mind, brain, and nervous system; and even how they garnered ever-wider contemporary purchase in cultures and societies are thus fascinating problems for historians of science and medicine. Such problems shed light on ethics, practices, controversies, and the uneasy social relations within those scientific and medical domains. But more to the point of this essay: they also account for the apparent epistemological weight now accorded "the neuro" in our contemporary moment. They thus illuminate in a rather different way why historians have suddenly discovered the value of "the neuro".