Biology subjects

Garibbo, M.

Publications and source records attributed to Garibbo, M..

3 recordsLinked to original sources

bioRxiv2026

Unifying error and reward action learning: a cerebello-basal ganglia theory

Learning depends on both reward- and error-based feedback, yet how the brain integrates these distinct signals to guide behaviour remains fundamentally unclear. Here, using a normative computational framework, we derive credit assignment rules for both reward-based learning (RBL) and error-based learning (EBL). In contrast to existing dual-policy accounts, our approach demonstrates that RBL and EBL updates can be reformulated into a shared action-gradient space that directly updates a single downstream policy. First, we map this action-gradient framework onto a systems-level account of coordinated interactions between the basal ganglia, cerebellum, and cortex. The model reproduces key behavioral features across both learning regimes, generates experimentally testable predictions, and provides a unified computational account of motor deficits observed in patients with cerebellar and basal ganglia disorders. Together, our work offers a normative, brain-wide framework for how distributed brain systems integrate reinforcement and error-driven feedback toward a common behavioral objective.

neuroscience↗

bioRxiv2026

ProtGPT3: an Open-source family of Promptable and Aligned Protein Language Models

Generative protein language models (pLMs) enable exploration of vast sequence spaces for protein design, but reliably controlling generation toward desired functional families remains challenging. While protein generation has broadly followed trends in NLP, two directions remain underexplored: alignment methods that optimize model behavior toward design objectives, and prompting-based control at inference time without fine-tuning. We introduce ProtGPT3, an open-source family of protein language models spanning 112M to 10B parameters and integrated with the Hugging Face ecosystem. The suite includes both single-sequence and multiple sequence alignment (MSA)-promptable models, enabling flexible conditioning for generation. Across model scales and protein families, we systematically compare supervised fine-tuning and few-shot prompting using homologous sequences. Analogous to how large language models (LLMs) are routinely aligned with user intent, we study post-training alignment in single-sequence models using sequence-complexity and structure-confidence metrics across the proteome. We find that alignment reduces low-complexity generations while preserving sequence diversity. Furthermore, we show that few-shot prompting is a competitive and more scalable alternative to supervised fine-tuning for controlled generation. In a low-data defluorinase case study, ProtGPT3-MSA achieved higher computational success rates than fine-tuned baselines and produced designs that were soluble and expressed following experimental validation. Finally, we explore the potential of inference-time compute in MSA models by introducing a homolog-based Feynman-Kac inference procedure for steering protein generation toward desired targets. All models are publicly available at https://huggingface.co/collections/AI4PD/protgpt3-family.

bioengineering↗

bioRxiv2025

Distinct roles of cortical layer 5 subtypes in associative learning

Adaptive behavior is critically dependent on associative learning, where environmental cues are linked with subsequent positive or negative outcomes. In mammals, primary neocortical sensory areas serve as pivotal nodes in this process, processing stimuli and distributing information to cortical and subcortical networks. Layer 5 (L5) of the cortex comprises two types of pyramidal projection neurons--intratelencephalic (IT) and extratelencephalic (ET) neurons--each with distinct downstream targets. Despite the crucial function of L5 as a main output node of the cortex, the specific contributions of these L5 neuronal subtypes to associative learning remain poorly understood. In the present study, by leveraging transgenic mouse lines, we distinguished IT and ET neurons in the primary somatosensory cortex and examined their roles in a whisker-based frequency-discrimination learning task. Longitudinal two-photon calcium imaging revealed distinct response characteristics between IT and ET neurons throughout learning. Interestingly, the activity of IT neurons hardly changed over the five days of learning, while the activity of ET neurons developed robustly. Furthermore, IT neurons appeared to show stimuli encoding from the beginning, whereas the ET neurons became increasingly responsive to stimuli associated with reward. Chemogenetic silencing of either IT or ET neurons both impaired learning, but in strikingly distinct ways, each associated with a different phase of learning. By modeling the response characteristics of IT and ET neurons using a reinforcement learning framework, we show that IT neurons primarily encode sensory stimuli, and their representations are critical for forming stimulus-reward associations. ET neurons instead represent the value of the stimulus, used for refining behavior. Thus, our results delineate the distinct roles of L5 IT and ET neurons, underscoring their integral and complementary contributions to associative learning.

neuroscience↗