中枢神経系疾患の治療は正常細胞（ニューロン）の機能維持を目的とするが，脳血管障害のように機能障害の原因が細胞の死滅に基づくことは多い．一方，脳腫瘍の治療においては薬物療法や放射線療法といった腫瘍細胞の死滅を目標とするものが大きな位置を占める．いずれの場合にも，細胞死の機序を理解することは各種病態や治療法の理解のうえで重要である．現在のところ最も研究の進んでいる細胞死の型はアポトーシスである．そのなかで重要な位置を占めるミトコンドリアにおける反応および抗アポトーシス因子について概要を紹介する．

Neuroscience 細胞死：最近の知見

Neuroscience has focused on the detailed implementation of computation, studying neural codes, dynamics and circuits. In machine learning, however, artificial neural networks tend to eschew precisely designed codes, dynamics or circuits in favor of brute force optimization of a cost function, often using simple and relatively uniform initial architectures. Two recent developments have emerged within machine learning that create an opportunity to connect these seemingly divergent perspectives. First, structured architectures are used, including dedicated systems for attention, recursion and various forms of short- and long-term memory storage. Second, cost functions and training procedures have become more complex and are varied across layers and over time. Here we think about the brain in terms of these ideas. We hypothesize that (1) the brain optimizes cost functions, (2) the cost functions are diverse and differ across brain locations and over development, and (3) optimization operates within a pre-structured architecture matched to the computational problems posed by behavior. In support of these hypotheses, we argue that a range of implementations of credit assignment through multiple layers of neurons are compatible with our current knowledge of neural circuitry, and that the brain's specialized systems can be interpreted as enabling efficient optimization for specific problem classes. Such a heterogeneously optimized system, enabled by a series of interacting cost functions, serves to make learning data-efficient and precisely targeted to the needs of the organism. We suggest directions by which neuroscience could seek to refine and test these hypotheses.

/pdf/toward-an-integration-of-deep-learning-and-neuroscience-ltox9rvhw6.pdf

Toward an Integration of Deep Learning and Neuroscience.

Toward an Integration of Deep Learning and Neuroscience

Neuroscience has focused on the detailed implementation of computation, studying neural codes, dynamics and circuits. In machine learning, however, artificial neural networks tend to eschew precisely designed codes, dynamics or circuits in favor of brute force optimization of a cost function, often using simple and relatively uniform initial architectures. Two recent developments have emerged within machine learning that create an opportunity to connect these seemingly divergent perspectives. First, structured architectures are used, including dedicated systems for attention, recursion and various forms of short- and long-term memory storage. Second, cost functions and training procedures have become more complex and are varied across layers and over time. Here we think about the brain in terms of these ideas. We hypothesize that (1) the brain optimizes cost functions, (2) these cost functions are diverse and differ across brain locations and over development, and (3) optimization operates within a pre-structured architecture matched to the computational problems posed by behavior. Such a heterogeneously optimized system, enabled by a series of interacting cost functions, serves to make learning data-efficient and precisely targeted to the needs of the organism. We suggest directions by which neuroscience could seek to refine and test these hypotheses.

/pdf/towards-an-integration-of-deep-learning-and-neuroscience-3586vtvx1w.pdf

Towards an integration of deep learning and neuroscience

The vertebrate control of locomotion involves all levels of the nervous system from cortex to the spinal cord. Here, we aim to cover all main aspects of this complex behavior, from the operation of...

https://journals.physiology.org/doi/pdf/10.1152/physrev.00015.2019

Current Principles of Motor Control, with Special Reference to Vertebrate Locomotion

Motor neurons (MNs) are neuronal cells located in the central nervous system (CNS) controlling a variety of downstream targets. This function infers the existence of MN subtypes matching the identity of the targets they innervate. To illustrate the mechanism involved in the generation of cellular diversity and the acquisition of specific identity, this review will focus on spinal motor neurons (SpMNs) that have been the core of significant work and discoveries during the last decades. SpMNs are responsible for the contraction of effector muscles in the periphery. Humans possess more than 500 different skeletal muscles capable to work in a precise time and space coordination to generate complex movements such as walking or grasping. To ensure such refined coordination, SpMNs must retain the identity of the muscle they innervate. Within the last two decades, scientists around the world have produced considerable efforts to elucidate several critical steps of SpMNs differentiation. During development, SpMNs emerge from dividing progenitor cells located in the medial portion of the ventral neural tube. MN identities are established by patterning cues working in cooperation with intrinsic sets of transcription factors. As the embryo develop, MNs further differentiate in a stepwise manner to form compact anatomical groups termed pools connecting to a unique muscle target. MN pools are not homogeneous and comprise subtypes according to the muscle fibers they innervate. This article aims to provide a global view of MN classification as well as an up-to-date review of the molecular mechanisms involved in the generation of SpMN diversity. Remaining conundrums will be discussed since a complete understanding of those mechanisms constitutes the foundation required for the elaboration of prospective MN regeneration therapies.

/pdf/motor-neurons-and-the-generation-of-spinal-motor-neuron-2b7utzoabm.pdf

Motor neurons and the generation of spinal motor neuron diversity

The spinal cord has the capacity to coordinate motor activities such as locomotion. Following spinal transection, functional activity can be regained, to a degree, following motor training. To identify microcircuits involved in this recovery, we studied a population of mouse spinal interneurons known to receive direct afferent inputs and project to intermediate and ventral regions of the spinal cord. We demonstrate that while dI3 interneurons are not necessary for normal locomotor activity, locomotor circuits rhythmically inhibit them and dI3 interneurons can activate these circuits. Removing dI3 interneurons from spinal microcircuits by eliminating their synaptic transmission left locomotion more or less unchanged, but abolished functional recovery, indicating that dI3 interneurons are a necessary cellular substrate for motor system plasticity following transection. We suggest that dI3 interneurons compare inputs from locomotor circuits with sensory afferent inputs to compute sensory prediction errors that then modify locomotor circuits to effect motor recovery.

/pdf/spinal-microcircuits-comprising-di3-interneurons-are-as0xsyoco2.pdf

Spinal microcircuits comprising dI3 interneurons are necessary for motor functional recovery following spinal cord transection.

Individual spinal motor neuron identities are specified in large part by the intrinsic repertoire of transcription factors expressed by undifferentiated progenitors and maturing neurons. It is shown here that the transcription factor AML1/Runx1 (Runx1) is expressed in selected spinal motor neuron subtypes after the onset of differentiation and is both necessary and sufficient to suppress interneuron-specific developmental programs and promote maintenance of motor neuron characteristics. These findings show an important role for Runx1 during the consolidation of selected spinal motor neuron identities. Moreover, they suggest a requirement for a persistent suppression of interneuron genes within maturing motor neurons.

https://www.pnas.org/content/pnas/105/17/6451.full.pdf

Suppression of interneuron programs and maintenance of selected spinal motor neuron fates by the transcription factor AML1/Runx1.

Spinal circuits for motor learning.

Kinematic gait parameters are highly sensitive measures of motor deficits and spinal cord injury in mice subjected to experimental autoimmune encephalomyelitis.

Nicolas Stifani

Papers

Motor neurons and the generation of spinal motor neuron diversity

Spinal microcircuits comprising dI3 interneurons are necessary for motor functional recovery following spinal cord transection.

Suppression of interneuron programs and maintenance of selected spinal motor neuron fates by the transcription factor AML1/Runx1.

Spinal circuits for motor learning.

Kinematic gait parameters are highly sensitive measures of motor deficits and spinal cord injury in mice subjected to experimental autoimmune encephalomyelitis.