Brain Inspired
A neuroscience-meets-AI podcast exploring the intersection of brain science and artificial intelligence. Host Paul Middlebrooks converses with neuroscientists, cognitive scientists, and AI researchers to uncover how insights about the brain can inform AI algorithms (and vice versa). In long-form discussions, they examine topics like neural coding, consciousness, neuromorphic computing, and deep learning, all aimed at understanding intelligence. An accessible yet deep show for those curious about how minds – biological or artificial – work, with new episodes posted regularly.
BI 234 Juan Gallego: The Neural Manifold Manifesto
In this episode, Juan Gallego discusses neural manifolds—mathematical structures that capture the coordinated activity of large neuron populations—and argues they are real, evolutionarily relevant objects with causal influence on behavior. He reviews evidence from monkey and mouse reaching tasks showing similar manifolds across species, and explains how low‑dimensional manifolds constrain learning, making adaptations within the manifold easy but changes outside it difficult. Gallego also describes his translational work using residual neural signals in spinal‑cord‑injury patients to infer intended movements and control computer simulations, aiming toward prosthetic control. Throughout, he balances enthusiasm for manifolds with caution about their limits in linking neural activity to cognition and mental phenomena.
BI 233 Tom Griffiths: The Laws of Thought
In this episode, Tom Griffiths discusses his new book, *The Laws of Thought*, which argues that cognition can be understood through three complementary pillars: logic, probability theory, and neural networks. He explains how logic provides deductive certainty, probability theory extends...
BI 231 Jaan Aru: Conscious AI? Not Even Close!
In this episode, Jaan Aru discusses how detailed biological mechanisms—such as dendritic integration and thalamocortical loops—might underpin subjective consciousness and how these insights could inform artificial intelligence. He argues that true machine consciousness is unlikely with current architectures, emphasizing the...
...
In this episode, Tomaso Poggio discusses extending Marr's three levels of analysis by adding learning as a fourth level, arguing that understanding intelligence requires both engineering breakthroughs and theoretical foundations—much like the era between Volta's battery and Maxwell's equations. He...
...
In this episode, philosophy of science professor Henk de Regt explains his theory of scientific understanding, arguing that true understanding requires the ability to generate theory‑based explanations and make qualitative predictions using the relevant scientific skills, rather than merely feeling a...
BI 196 Cristina Savin and Tim Vogels with Gaute Einevoll and Mikkel Lepperød
The post recaps episode BI 196 of the *Brain Inspired* podcast, where hosts Gaute Einevoll and a guest discuss Neuro‑AI with researchers Cristina Savin and Tim Vogels. Savin describes using recurrent neural networks to model learning and behavior, while Vogels explains how AI‑driven optimization is...
BI 195 Ken Harris and Andreas Tolias with Gaute Einevoll and Mikkel Lepperød
In this episode of Brain Inspired, hosts Paul and Gaute Einevoll share recordings from a Norwegian Neuro‑AI workshop, featuring conversations with neuroscientists Ken Harris and Andreas Tolias. Harris discusses his ultra‑high‑density recordings of thousands of neurons and the challenges of...
BI 191 Damian Kelty-Stephen: Fractal Turbulent Cascading Intelligence
In this Brain Inspired episode, experimental psychologist Damian Kelty‑Stephen critiques the dominant computer‑metaphor of the brain and proposes that fractal, cascade, and turbulence dynamics—rooted in ecological psychology—better explain intelligence and behavior across scales. He traces his academic journey from developmental...
BI 190 Luis Favela: The Ecological Brain
The post introduces Luis Favela’s new book *The Ecological Brain*, which puts forward the NeuroEcological Nexus Theory (NExT) to unify neuroscience, ecological psychology, and the body‑environment system through low‑dimensional neural, bodily, and environmental dynamics and manifold mathematics. Favela argues that...
BI 189 Joshua Vogelstein: Connectomes and Prospective Learning
In this episode Joshua Vogelstein discusses two core themes: the creation of the world’s largest whole‑brain connectome for the fruit fly and his team’s concept of prospective learning, which contrasts with the dominant retrospective learning in AI. He explains how...
