Follow developments in nootropics and brain research with their sources and limitations. We explain what a new finding says, and what it cannot establish.
In an Ecuadorian randomized trial of 215 pregnancies, a food-diversity program was associated with growth differences in several fetal brain structures; the study measured anatomy, not later cognition.
A 24-week randomized trial in 87 people with early cognitive impairment found lower pTau181 and selective visuospatial signals, but no MMSE, verbal-memory or executive-score difference.
In 31 young adults, EEG during word–object learning linked stronger theta activity with later sleep-wave coupling and memory retention after a nap; the association does not establish causation.
In an fMRI experiment, retrieval prompts shaped later neural patterns for learned image–word pairs, even though participants performed similarly on the memory test.
A 70-person randomized Parkinson trial found short-term walking-speed and later freezing-of-gait signals from stimulation plus treadmill training, but no sustained overall advantage.
In a randomized trial of young men, six hours of simulated hypoxia at 4,500 m did not alter psychobiological measures, and glutamine supplementation did not produce a detectable benefit.
An August NIH report highlighted altered DNA contacts and gene activity in postmortem prefrontal cells from Alzheimer’s donors; the study does not establish causality.
A 12-week peanut trial in 27 older adults with memory complaints found no significant group differences in cognition, brain activation or adjusted biomarkers.
Manipulating wake-responsive neurons in mice altered NREM sleep and sleep pressure, identifying candidate circuits for sleep regulation without showing that reduced sleep is safe.
NIH reported that stem-cell-derived brain organoids matured over nearly six years, with molecular age tracking elapsed time; these are laboratory models.
During a delayed-match task, mice reused prefrontal and parietal neural subspaces for stimulus processing and memory maintenance. Data-constrained network models supported a modular account, but the result concerns one task and does not prove human cognitive modules.