Scribd
Upload
42 views

CompneuroFinalDiscussion 2019

The document summarizes a report from the NIH called "Brain 2025: A Scientific Vision" from 2014. The report outlines 7 goals for neuroscience research over the next decade: 1) Identifying all brain cell types, 2) Mapping neural circuits at all scales, 3) Developing methods to monitor widespread neural activity, 4) Developing tools to causally link brain activity to behavior, 5) Developing theoretical and data analysis tools, 6) Advancing human neuroscience research, and 7) Integrating new approaches to understand cognition, perception, and behavior. The overarching goal is to develop a comprehensive understanding of the brain and mental processes by achieving these scientific objectives.

Uploaded by

cumte simti
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
42 views

CompneuroFinalDiscussion 2019

The document summarizes a report from the NIH called "Brain 2025: A Scientific Vision" from 2014. The report outlines 7 goals for neuroscience research over the next decade: 1) Identifying all brain cell types, 2) Mapping neural circuits at all scales, 3) Developing methods to monitor widespread neural activity, 4) Developing tools to causally link brain activity to behavior, 5) Developing theoretical and data analysis tools, 6) Advancing human neuroscience research, and 7) Integrating new approaches to understand cognition, perception, and behavior. The overarching goal is to develop a comprehensive understanding of the brain and mental processes by achieving these scientific objectives.

Uploaded by

cumte simti
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

Computational Neuroscience

2019

Final discussion
Odelia Schwartz

1
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#1. Discovering diversity: Identify and provide experimental


access to the different brain cell types to determine their roles in
health and disease. It is within reach to characterize all cell types in
the nervous system, and to develop tools to record, mark, and
manipulate these precisely defined neurons in the living brain. We
envision an integrated, systematic census of neuronal and glial cell
types, and new genetic and non-­‐genetic tools to deliver genes,
proteins, and chemicals to cells of interest in non-­‐human animals
and in humans.

2
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#2. Maps at multiple scales: Generate circuit diagrams that vary


in resolution from synapses to the whole brain. It is increasingly
possible to map connected neurons in local circuits and distributed
brain systems, enabling an understanding of the relationship
between neuronal structure and function. We envision improved
technologies—faster, less expensive, scalable— for anatomic
reconstruction of neural circuits at all scales, from non-­‐invasive
whole human brain imaging to dense reconstruction of synaptic
inputs and outputs at the subcellular level.

3
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#3. The brain in action: Produce a dynamic picture of the


functioning brain by developing and applying improved methods for
large-­‐scale monitoring of neural activity. We should seize the
challenge of recording dynamic neuronal activity from complete
neural networks, over long periods, in all areas of the brain. There
are promising opportunities both for improving existing
technologies and for developing entirely new technologies for
neuronal recording, including methods based on electrodes, optics,
molecular genetics, and nanoscience, and encompassing different
facets of brain activity.

4
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#4. Demonstrating causality: Link brain activity to behavior with


precise interventional tools that change neural circuit dynamics. By
directly activating and inhibiting populations of neurons,
neuroscience is progressing from observation to causation, and
much more is possible. To enable the immense potential of circuit
manipulation, a new generation of tools for optogenetics,
chemogenetics, and biochemical and electromagnetic modulation
should be developed for use in animals and eventually in human
patients.

5
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#5. Identifying fundamental principles: Produce conceptual


foundations for understanding the biological basis of mental
processes through development of new theoretical and data
analysis tools. Rigorous theory, modeling, and statistics are
advancing our understanding of complex, nonlinear brain functions
where human intuition fails. New kinds of data are accruing at
increasing rates, mandating new methods of data analysis and
interpretation. To enable progress in theory and data analysis, we
must foster collaborations between experimentalists and scientists
from statistics, physics, mathematics, engineering, and computer
science.

6
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#6. Advancing human neuroscience: Develop innovative


technologies to understand the human brain and treat its disorders;
create and support integrated human brain research networks.
Consenting humans who are undergoing diagnostic brain
monitoring, or receiving neurotechnology for clinical applications,
provide an extraordinary opportunity for scientific research. This
setting enables research on human brain function, the mechanisms
of human brain disorders, the effect of therapy, and the value of
diagnostics. Meeting this opportunity requires closely integrated
research teams performing according to the highest ethical
standards of clinical care and research. New mechanisms are
needed to maximize the collection of this priceless information and
ensure that it benefits people with brain disorders.
7
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

#7. From BRAIN Initiative to the brain: Integrate new


technological and conceptual approaches produced in Goals #1-­‐6 to
discover how dynamic patterns of neural activity are transformed
into cognition, emotion, perception, and action in health and
disease. The most important outcome of the BRAIN Initiative will be
a comprehensive, mechanistic understanding of mental function
that emerges from synergistic application of the new technologies
and conceptual structures developed under the BRAIN Initiative.

The overarching vision of the BRAIN Initiative is best


captured by Goal #7—combining these approaches into a
single, integrated science of cells, circuits, brain, and
behavior.
8
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

Cross boundaries in interdisciplinary collaborations. No single


researcher or discovery will solve the brain’s mysteries. The most
exciting approaches will bridge fields, linking experiment to theory,
biology to engineering, tool development to experimental
application, human neuroscience to non-­‐human models, and more,
in innovative ways.

9
From the NIH web site:
Committee report: Brain 2025: A Scientific Vision
(from 2014)

Theory, Modeling, and Statistics Will Be Essential to


Understanding the Brain

Ideally, theorists and statisticians should be involved in


experimental design and data acquisition, not just recruited at the
step of data interpretation.

10
Lots of recent interest from industry

11
Google Brain

“We are a machine intelligence team focused on deep


learning. We advance the state of the art in order to
have a positive impact on the world. We achieve this
goal by focusing on highly flexible models that learn
their own features, end-to-end, and make efficient use
of data and computation. This approach fits into the
broader Deep Learning subfield of ML and ensures
our work will ultimately make a difference for
problems of practical importance.”

12
Computational neuroscience

Deep Mind: solve intelligence.


Use it to make the world a better place

“We joined forces with Google in order to turbo-


charge our mission. The algorithms we build are
capable of learning for themselves directly from raw
experience or data, and are general in that they can
perform well across a wide variety of tasks straight out
of the box. Our world-class team consists of many
renowned experts in their respective fields, including
but not limited to deep neural networks, reinforcement
learning and systems neuroscience-inspired models.”
13
Computational neuroscience

IBM Research: Cognitive computing:

“The Cognitive Era: By any measure, 2015 has been a


landmark year for the discussion around artificial
intelligence and its potential impact on business and
society. Be part of the conversation as we explore a
fascinating and diverse set of issues related to the
powerful cognitive technologies that are emerging to
augment human capacity and understanding.”

14
Computational neuroscience
C O M M E N TA RY
F O C U S O N N E U R A L C O M P U TAT I O N A N D T H E O R Y

Conceptual and technical advances define


a key moment for theoretical neuroscience
Anne K Churchland & L F Abbott

Theoretical approaches have long shaped neuroscience, but current needs for theory are elevated and prospects for advancement
are bright. Advances in measuring and manipulating neurons demand new models and analyses to guide interpretation. Advances
in theoretical neuroscience offer new insights into how signals evolve across areas and new approaches for connecting population
activity with behavior. These advances point to a global understanding of brain function based on a hybrid of diverse approaches.

Theoretical approaches have a long history complex data into forms that summarize the This approach differs critically from tradi-
of contributing to neuroscience research, but results in a more compact and understandable tional averaging in that neurons contribute in
never before has the need for them been so way. Third, using modeling to link the results to varying degrees, or even negatively (Fig. 1),
high nor the prospects for advancement so underlying mechanisms and overlying princi- to the extracted signal. Simple classifiers are
great. The explosion in technologies available ples. The perspectives and reviews in this issue becoming a standard tool for interpreting neu-
for measuring and manipulating neurons has primarily address the third, and some of the ral data10,11. More complex classifiers, such
created a call for analysis techniques that are second, stage, surveying new developments and as those with multiple processing layers, are
scalable to extremely large data sets, that take modeling-based insights in topics ranging from rapidly being developed in both neuroscience
and industry12. One example is the hierarchi-
15 Churchland, A.K. and Abbott, L.F. (2016)
into account the heterogeneity of neurons,
and that can predict and interpret the effects
understanding and interpreting network spik-
ing activity1–3, exploring visual processing4 and cal convolutional neural network (HCNN),
of complex manipulations of activity. The memory5, and studying the representation and which is based on ideas developed in stud-
growing importance of theory is also driven computation of probability6, to investigations ies of the visual system. As discussed in this
by developments in theoretical neuroscience of higher level cognition and mental illness7. issue, HCNNs can be used to interpret activ-
itself, advances that expand the reach of theo- In addition, major advances in the other stages ity at various stages along the primate visual
retical approaches and extend their ability to of analysis have driven the entire program to pathway during object recognition4.
offer insight into long-standing puzzles. In evolve considerably in recent years. Two such Second, new challenges for data analysis
Computational neuroscience

“In the coming years, we will obtain enormous quantities of


behavioral, recording (both electrical and optical),
connectomic, gene expression and other forms of data.
Obtaining deep understanding from this onslaught will
require, in addition to the skillful and creative application of
experimental technologies, substantial advances in data
analysis methods and intense application of theoretic
concepts and models.”

16
Fields evolving include …

• Computational Cognitive Neuroscience (new


conference combining AI, Cognitive Science,
Neuroscience)
• Decision making and reinforcement learning.
“Perceptual Decision-Making: A Field in the Midst of a
Transformation” (Najafi, Churchland 2018)
• Computational Psychiatry
• Vision and learning …
• New approaches and analyses for making sense of large scale data

17
Related research areas At UM

• Computational Neuroscience
• Neural Engineering and Brain machine Interfaces
• Machine learning
• Data science
• Large-scale fMRI
• Technology such as optogenetics
• Neuroscience / Biology
• Robotics

18
19

You might also like