@NootropicsDAO_

Gmind everyone. Thanks for all the support during the launch announcement. We will be releasing more information and details about the app we built, the AI models we trained, the utility of the product, and a demo really soon. Stay tunned! For more info: https://t.co/Vb7HlSDpHL

How well can Neurolab predict molecular potency? To understand how a drug may affect the brain, we first ask which molecular targets it acts on—and at what concentrations. We tested one part of that challenge: predicting potency for specified targets, using 30 MoleculeACE datasets and 9,802 test observations. The benchmark measures two things: • Overall error: how closely predictions match measured potency across the test observations. • Activity-cliff error: performance on similar molecules with very different potency—cases where small structural changes can matter substantially. We rank these separately, giving each dataset equal weight. Lower error is better. The results: overall / activity-cliff RMSE: • Neurolab 1: 0.6562 / 0.7458 • Neurolab 2: 0.6559 / 0.7576 • CheMeleon, two-run mean: 0.6605 / 0.7438 Neurolab 2 recorded the lowest overall error among the 29 entries compared. Neurolab 1 was second overall and second on activity cliffs. CheMeleon led the cliff comparison. These are errors in log10 potency units—not accuracy percentages. So, which model should we use? Cliff error was our primary metric. Neurolab 2’s tiny overall advantage does not automatically make it the better choice. Likewise, CheMeleon’s slightly lower cliff error does not settle the deployment decision. These close results do not establish statistical superiority. The complete comparison retains all 24 published reference configurations, four Neurolab variants and CheMeleon’s two-run mean. It tests molecular-potency prediction, not mechanisms of action, clinical outcomes or the full Neurolab system.

ATTENTION: DeSci Community We're proud to announce that the Neurolab Ignition Sale is going LIVE between September 15th - September 17th on the @BioProtocol launchpad! Explore the NeuroLab Launch Command Center, read the official project description on the Bio launchpad, and get ready to participate: https://t.co/Vb7HlSDpHL The moment we've been looking forward is almost here. From Nootropics DAO to NeuroLab, the ambition is clear: bring cutting-edge, AI tools to enable community-led CNS drug discovery to the people asking the next generation of fundamental questions about the human brain. Now we're inviting the DeSci, biotech and crypto communities to get to know NeuroLab and take part in this milestone. What exactly is Neurolab? Built to be DeSci's first true novel research workspace centered around the human brain, the NeuroLab application uses a proprietary collection of graph neural networks, deep learning, and neurosymbolic reasoning to predict the cognitive, molecular, and functional effects of any known or novel chemical compound on the human brain. From compounds to molecular targets and all the way up through cortical and subcortical regions, Neurolab's multiscale reasoning architecture aims to help researchers investigate ideas, examine the evidence behind them, and decide which questions deserve a closer look. Over the coming days, we'll share the product workflow, the research behind our approach, our planned roadmap and how to participate in the Ignition Sale. Thank you to everyone who has followed, challenged and supported the work. We've been looking forward to sharing this moment with you and setting a new standard in decentralized science products for professional researchers and rigorous citizen scientists alike 🤝

Interesting bromantane made it in there! Many people in the nootropics community synergize bromantane with tyrosine (which also made it onto the list, on its own). Underrated "hack": Customize any nootropics you're taking to the specific nature of the task at hand 🧠

✍️ https://t.co/2uqqsSplPs