Positioning spontaneous activity as ‘Adhesive Dots’: Lessons from AI for data integration in neuroscience In the previous article, I argued that advancing data integration in neuroscience requires incorporating resting-state spontaneous activity into each experiment, framing it as ‘adhesive dots.’ Here, I extend that discussion by drawing strategic lessons from the success of large language models (LLMs) and by concretizing the earlier claims from the perspective of data. The worldwide construction of data centers illustrates how AI development has advanced through scaling – expanding data volume, model size, and computational resources. LLM performance improves according to power-law scaling when all three expand together. (1) Furthermore, scaling model size and dataset size in tandem has been shown to be near-optimal. (2)
Masanori Shimono (2025) studied this question.