Computational Neuroscience and Neuroplasticity: Implications for Christian Belief

Transcription

1 Computational Neuroscience and Neuroplasticity: Implications for Christian Belief DANIEL DORMAN AMERICAN SCIENTIFIC AFFILIATE ANNUAL CONFERENCE, JULY 2016

2 Big Questions Our human intelligence is based on computational processes that we are learning to understand. We will ultimately multiply our intellectual powers by applying and extending the methods of human intelligence using the vastly greater capacity of nonbiological computation. Ray Kurzweil, The Singularity is Near: When Humans Transcend Biology

3 Big Questions Can a computational simulation of a brain suffer?. If the answer is yes, it then follows that we should consider extending the same protections against suffering that are given to subjects of experimental research. (Bancroft, Ethical Aspects of Computational Neuroscience, Neuroethics, 2013)

4 Big Questions Here is an idea that would make Descartes blanch. Apparently, there are scientists who believe that at some point fairly soon we will be able to upload our minds to computers, freeing ourselves from our bodies, being, therefore, immortal. I suppose they will program in the virtual experience of taking the uploaded dog on a walk to the virtual park, through the rain on randomly assorted virtual days adjusted to reflect prevailing weather patterns in some selected place and season. These immortals would at last be free of the thousand natural shocks that flesh is heir to, and with them no doubt of all urgent reflection on what we are and what we mean. I can t help imagining that, given the sterility of it all, the sullens would set in and these uploaded minds would do what many of their creators do devise ingenious viruses, spy on one another, refine resentments, contrive schemes to dupe the mortals. Then physical reality, let us say a great solar storm, the impact of a meteor, even a major war, would sweep them all away, making the always necessary point that, for our purposes, the physical is not to be transcended. In any case, an uploaded mind would be as void of soul as a cryogenically frozen body. We know this intuitively.

5 Is such a vision theoretically possible, grounded in current progress in the field of computational neuroscience? Desirable from a Christian and Humanist perspective?

6 Outline Conceptual Foundations for Computational Neuroscience Current progress and perspectives in the field My work: Computational Neuroplasticity Framework for Understanding Computation Neuroscience

7 Conceptual Foundations Computational Description and Prediction Computational Neurons/Brains Theory of Mind Compute

8 Progress and Perspectives A Comparison of Two Approaches Top Down, Functional Approach SPAUN Bottom up, biological approach The Blue Brain Project

9 SPAUN A Behaving Brain Simulation 2.5 Million Neuron Brain Model Approximates known brain circuits Sensing eye Robotic Arm Engineered to perform specific functions Eliasmith, Chris, et al. "A large-scale model of the functioning brain." Science (2012):

10

11 Blue Brain Project Biologically Driven Cortical Column Model Experimental Reconstruction Detailed model Emergent neural activity patterns Markram et al. Reconstruction and Simulation of Neocortical Microcircuitry. Cell 2015;163(2):

12

13 Summary of Contrasting Models

14 My Research: Single Cell Models and Subcellular Mechanisms of Synaptic Plasticity Kotaleski & Blackwell,

15 Computational Neuroplasticity Phenomenon Neuroplasticity of Striatum Implications Computational Model wikimedia Predictions Experimental Data

16 The striatum is the input nucleus to the basal ganglia and is associated with reward-driven action and learning Cognitive associations Action selection Goal-directed & reinforcement learning Habit formation Scholarpedia.org Kravitz and Kreitzer, 2012 How does a single striatal neuron participate in learning and action? 16

17 Supralinear Response to Cooperative Synaptic Input

18 Presynaptic Activity Patterns Trigger Biochemical Signalling Pathways that Alter the Physiology and Structure of the Neuron and Modify Genetic Expression at the Nucleus Bhalla US. Molecular computation in neurons: a modeling perspective. Curr. Opin. Neurobiol. 2014

19 Interpretive Framework Genetic Expression Cellular Activity Network Activity Embodied Action Situated in Environment

20 Level 1 computational methods enable description and explanation of how the physical substrate of the brain may be said to compute, to process neural information (level 2). Does this then support level 3, that (a) the mind is strictly computational, in an abstracted sense that makes the brain accidental, such that any computing substrate could have a mind, or (b) such that level 1 computational simulations could possess an internal, subjective, conscious self? I argue that neuroscience challenges level 3, because in the brain, information and computation are physically, neuronally, embodied, whereas in a simulation there are multiple levels of abstraction: We describe physical neurons in abstract mathematical terms, and we then translate the mathematical equations to computer algorithms which yield numbers; it is only our understanding and interpretation of those numbers that render the results meaningful, the computer could care less. However, new physical conceptions of computers are being designed to parallel the biology of the brain. If brain simulations can be physically embodied in such computers, without abstraction, we may be surprised.

21 SPAUN A Behaving Brain Simulation 1. Eliasmith C, Trujillo O. The use and abuse of large-scale brain models. Curr. Opin. Neurobiol. 2014;25:1-6. doi: /j.conb

22 1. Kaplan DM. Explanation and description in computational neuroscience. Synthese 2011;183(3): doi: / s

23 1. Piccinini G, Shagrir O. Foundations of computational neuroscience. Curr. Opin. Neurobiol. 2014;25: doi: / j.conb

24 1. Critical Neuroscience.

25 1. Markram H, Muller E, Ramaswamy S, et al. Reconstruction and Simulation of Neocortical Microcircuitry. Cell 2015;163(2): doi: /j.cell