He argues that our brain’s computations are of a completely different nature than any artificial intelligence because they take place across many spatial and temporal scales and are inextricably…

The rapid advances in the capabilities of Large Language Models (LLMs) have galvanised public and scientific debates over whether artificial systems m…

The rapid advances in the capabilities of Large Language Models (LLMs) have galvanised public and scientific debates over whether artificial systems m…

Our understanding of the neural basis of consciousness is mostly restricted to large-scale brain activity patterns as measured by methods such as functional magnetic resonance imaging (fMRI) and magneto/electro-encephalography (M/EEG). In contrast, we lack even basic understanding of circuit-level mechanisms supporting consciousness – particularly in humans – despite the fundamental role that such mechanisms likely play in instantiating larger-scale brain activity patterns supporting conscious states and contents. Here, we review what progress has been made on circuit-level theories of consciousness (e.g., apical amplification theory, dendritic integration theory) and argue that such theories can be tested in humans using recently developed, state-of-the-art methods. Doing so will further facilitate translation of consciousness science into clinical settings and strengthen the bridge between circuit- and network-level theories of consciousness.

Artificial neural networks are becoming more advanced and human-like in detail and behavior. The notion that machines mimicking human brain computations migh...