BI 220 Michael Breakspear and Mac Shine: Dynamic Systems from Neurons to Brains
Support the show to get full episodes, full archive, and join the Discord community.
The Transmitter is an online publication that aims to deliver useful information, insights and tools to build bridges across neuroscience and advance research. Visit thetransmitter.org to explore the latest neuroscience news and perspectives, written by journalists and scientists.
Read more about our partnership: https://www.thetransmitter.org/partners/
Sign up for the “Brain Inspired” email alerts to be notified every time a new “Brain Inspired” episode is released: https://www.thetransmitter.org/newsletters/
To explore more neuroscience news and perspectives, visit thetransmitter.org.
What changes and what stays the same as you scale from single neurons up to local populations of neurons up to whole brains? How tuning parameters like the gain in some neural populations affects the dynamical and computational properties of the rest of the system.
Those are the main questions my guests today discuss. Michael Breakspear is a professor of Systems Neuroscience and runs the Systems Neuroscience Group at the University of Newcastle in Australia. Mac Shine is back, he was here a few years ago. Mac runs the Shine Lab at the University of Sidney in Australia.
Michael and Mac have been collaborating on the questions I mentioned above, using a systems approach to studying brains and cognition. The short summary of what they discovered in their first collaboration is that turning up or down the gain across broad networks of neurons in the brain affects integration - working together - and segregation - working apart. They map this gain modulation on to the ascending arousal pathway, in which the locus coeruleus projects widely throughout the brain distributing noradrenaline. At a certain sweet spot of gain, integration and segregation are balanced near a bifurcation point, near criticality, which maximizes properties that are good for cognition.
In their recent collaboration, they used a coarse graining procedure inspired by physics to study the collective dynamics of various sizes of neural populations, going from single neurons to large populations of neurons. Here they found that despite different coding properties at different scales, there are also scale-free properties that suggest neural populations of all sizes, from single neurons to brains, can do cognitive stuff useful for the organism. And they found this is a conserved property across many different species, suggesting it's a universal principle of brain dynamics in general.
So we discuss all that, but to get there we talk about what a systems approach to neuroscience is, how systems neuroscience has changed over the years, and how it has inspired the questions Michael and Mac ask.
Breakspear: Systems Neuroscience Group.
@DrBreaky.
Shine: Shine Lab.
@jmacshine.
Related papers
Dynamic models of large-scale brain activity
Metastable brain waves
The modulation of neural gain facilitates a transition between functional segregation and integration in the brain
Multiscale Organization of Neuronal Activity Unifies Scale-Dependent Theories of Brain Function.
The brain that controls itself.
Metastability demystified — the foundational past, the pragmatic present and the promising future.
Generation of surrogate brain maps preserving spatial autocorrelation through random rotation of geometric eigenmodes.
Related episodes
BI 212 John Beggs: Why Brains Seek the Edge of Chaos
BI 216 Woodrow Shew and Keith Hengen: The Nature of Brain Criticality
BI 121 Mac Shine: Systems Neurobiology
0:00 - Intro
4:28 - Neuroscience vs neurobiology
8:01 - Systems approach
26:52 - Physics for neuroscience
33:15 - Gain and bifurcation: earliest collaboration
55:32 - Multiscale organization
1:17:54 - Roadblocks
--------
1:25:05
--------
1:25:05
BI 219 Xaq Pitkow: Principles and Constraints of Cognition
Support the show to get full episodes, full archive, and join the Discord community.
The Transmitter is an online publication that aims to deliver useful information, insights and tools to build bridges across neuroscience and advance research. Visit thetransmitter.org to explore the latest neuroscience news and perspectives, written by journalists and scientists.
Read more about our partnership.
Sign up for Brain Inspired email alerts to be notified every time a new Brain Inspired episode is released.
To explore more neuroscience news and perspectives, visit thetransmitter.org.
Xaq Pitkow runs the Lab for the Algorithmic Brain at Carnegie Mellon University. The main theme of our discussion is how Xaq approaches his research into cognition by way of principles, from which his questions and models and methods spring forth. We discuss those principles, and In that light, we discuss some of his specific lines of work and ideas on the theoretical side of trying understand and explain a slew of cognitive processes. A few of the specifics we discuss are:
How when we present tasks for organisms to solve, they use strategies that are suboptimal relative to the task, but nearly optimal relative to their beliefs about what they need to do - something Xaq calls inverse rational control.
Probabilistic graph networks.
How brains use probabilities to compute.
A new ecological neuroscience project Xaq has started with multiple collaborators.
LAB: Lab for the Algorithmic Brain.
Related papers
How does the brain compute with probabilities?
Rational thoughts in neural codes.
Control when confidence is costly
Generalization of graph network inferences in higher-order graphical models.
Attention when you need.
0:00 - Intro
3:57 - Xaq's approach
8:28 - Inverse rational control
19:19 - Space of input-output functions
24:48 - Cognition for cognition
27:35 - Theory vs. experiment
40:32 - How does the brain compute with probabilities?
1:03:57 - Normative vs kludge
1:07:44 - Ecological neuroscience
1:20:47 - Representations
1:29:34 - Current projects
1:36:04 - Need a synaptome
1:42:20 - Across scales
--------
1:47:11
--------
1:47:11
BI 218 Chris Rozell: Brain Stimulation and AI for Mental Disorders
Support the show to get full episodes, full archive, and join the Discord community.
We are in an exciting time in the cross-fertilization of the neurotech industry and the cognitive sciences. My guest today is Chris Rozell, who sits in that space that connects neurotech and brain research. Chris runs the Structured Information for Precision Neuroengineering Lab at Georgia Tech University, and he was just named the inaugural director of Georgia Tech’s Institute for Neuroscience, Neurotechnology, and Society. I think this is the first time on brain inspired we've discussed stimulating brains to treat mental disorders. I think. Today we talk about Chris's work establishing a biomarker from brain recordings of patients with treatment resistant depression, a specific form of depression. These are patients who have deep brain stimulation electrodes implanted in an effort to treat their depression. Chris and his team used that stimulation in conjunction with brain recordings and machine learning tools to predict how effective the treatment will be under what circumstances, and so on, to help psychiatrists better treat their patients. We'll get into the details and surrounding issues. Toward the end we also talk about Chris's unique background and path and approach, and why he thinks interdisciplinary research is so important. He's one of the most genuinely well intentioned people I've met, and I hope you're inspired by his research and his story.
Structured Information for Precision Neuroengineering Lab.
Twitter: @crozSciTech.
Related papers
Cingulate dynamics track depression recovery with deep brain stimulation.
Story Collider: Wired Lives
0:00 - Intro
3:20 - Overview of the study
17:11 - Closed and open loop stimulation
19:34 - Predicting recovery
28:45 - Control knob for treatment
39:04 - Historical and modern brain stimulation
49:07 - Treatment resistant depression
53:44 - Control nodes complex systems
1:01:06 - Explainable generative AI for a biomarker
1:16:40 - Where are we and what are the obstacles?
1:21:32 - Interface Neuro
1:24:55 - Why Chris cares
Read the transcript.
--------
1:46:39
--------
1:46:39
BI 217 Jennifer Prendki: Consciousness, Life, AI, and Quantum Physics
Support the show to get full episodes, full archive, and join the Discord community.
The Transmitter is an online publication that aims to deliver useful information, insights and tools to build bridges across neuroscience and advance research. Visit thetransmitter.org to explore the latest neuroscience news and perspectives, written by journalists and scientists.
Read more about our partnership.
Sign up for Brain Inspired email alerts to be notified every time a new Brain Inspired episode is released.
To explore more neuroscience news and perspectives, visit thetransmitter.org.
Do AI engineers need to emulate some processes and features found only in living organisms at the moment, like how brains are inextricably integrated with bodies? Is consciousness necessary for AI entities if we want them to play nice with us? Is quantum physics part of that story, or a key part, or the key part? Jennifer Prendki believes if we continue to scale AI, it will get us more of the same of what we have today, and that we should look to biology, life, and possibly consciousness to enhance AI. Jennifer is a former particle physicist turned entrepreneur and AI expert, focusing on curating the right kinds and forms of data to train AI, and in that vein she led those efforts at Deepmind on the foundation models ubiquitous in our lives now.
I was curious why someone with that background would come to the conclusion that AI needs inspiration from life, biology, and consciousness to move forward gracefully, and that it would be useful to better understand those processes in ourselves before trying to build what some people call AGI, whatever that is. Her perspective is a rarity among her cohorts, which we also discuss. And get this: she's interested in these topics because she cares about what happens to the planet and to us as a species. Perhaps also a rarity among those charging ahead to dominate profits and win the race
Jennifer's website: Quantum of Data.
The blog posts we discuss:
The Myth of Emergence
Embodiment & Sentience: Why the Body still Matters
The Architecture of Synthetic Consciousness
On Time and Consciousness
Superalignment and the Question of AI Personhood.
Read the transcript.
0:00 - Intro
3:25 - Jennifer's background
13:10 - Consciousness
16:38 - Life and consciousness
23:16 - Superalignment
40:11 - Quantum
1:04:45 - Wetware and biological mimicry
1:15:03 - Neural interfaces
1:16:48 - AI ethics
1:2:35 - AI models are not models
1:27:13 - What scaling will get us
1:39:53 - Current roadblocks
1:43:19 - Philosophy
--------
1:48:53
--------
1:48:53
BI 216 Woodrow Shew and Keith Hengen: The Nature of Brain Criticality
Support the show to get full episodes, full archive, and join the Discord community.
The Transmitter is an online publication that aims to deliver useful information, insights and tools to build bridges across neuroscience and advance research. Visit thetransmitter.org to explore the latest neuroscience news and perspectives, written by journalists and scientists.
Read more about our partnership.
Sign up for Brain Inspired email alerts to be notified every time a new Brain Inspired episode is released.
To explore more neuroscience news and perspectives, visit thetransmitter.org.
A few episodes ago, episode 212, I conversed with John Beggs about how criticality might be an important dynamic regime of brain function to optimize our cognition and behavior. Today we continue and extend that exploration with a few other folks in the criticality world.
Woodrow Shew is a professor and runs the Shew Lab at the University of Arkansas. Keith Hengen is an associate professor and runs the Hengen Lab at Washington University in St. Louis Missouri. Together, they are Hengen and Shew on a recent review paper in Neuron, titled Is criticality a unified setpoint of brain function? In the review they argue that criticality is a kind of homeostatic goal of neural activity, describing multiple properties and signatures of criticality, they discuss multiple testable predictions of their thesis, and they address the historical and current controversies surrounding criticality in the brain, surveying what Woody thinks is all the past studies on criticality, which is over 300. And they offer a account of why many of these past studies did not find criticality, but looking through a modern lens they most likely would. We discuss some of the topics in their paper, but we also dance around their current thoughts about things like the nature and implications of being nearer and farther from critical dynamics, the relation between criticality and neural manifolds, and a lot more. You get to experience Woody and Keith thinking in real time about these things, which I hope you appreciate.
Shew Lab. @ShewLab
Hengen Lab.
Is criticality a unified setpoint of brain function?
Read the transcript.
0:00 - Intro
3:41 - Collaborating
6:22 - Criticality community
14:47 - Tasks vs. Naturalistic
20:50 - Nature of criticality
25:47 - Deviating from criticality
33:45 - Sleep for criticality
38:41 - Neuromodulation for criticality
40:45 - Criticality Definition part 1: scale invariance
43:14 - Criticality Definition part 2: At a boundary
51:56 - New method to assess criticality
56:12 - Types of criticality
1:02:23 - Value of criticality versus other metrics
1:15:21 - Manifolds and criticality
1:26:06 - Current challenges
Neuroscience and artificial intelligence work better together. Brain inspired is a celebration and exploration of the ideas driving our progress to understand intelligence. I interview experts about their work at the interface of neuroscience, artificial intelligence, cognitive science, philosophy, psychology, and more: the symbiosis of these overlapping fields, how they inform each other, where they differ, what the past brought us, and what the future brings. Topics include computational neuroscience, supervised machine learning, unsupervised learning, reinforcement learning, deep learning, convolutional and recurrent neural networks, decision-making science, AI agents, backpropagation, credit assignment, neuroengineering, neuromorphics, emergence, philosophy of mind, consciousness, general AI, spiking neural networks, data science, and a lot more. The podcast is not produced for a general audience. Instead, it aims to educate, challenge, inspire, and hopefully entertain those interested in learning more about neuroscience and AI.
Brasil