Science

Current Picks

One Uncorrected fMRI Head Motion Threshold Shifts a Whole-Brain Functional Connectivity Map

By Jonas Eriksen / Jun 12, 2026

A 0.5 mm change in fMRI head motion threshold can rewire whole-brain connectivity maps, creating false circuits. The problem is rooted in research incentives and costly scanner time.

A Single Unfunded Precision Mirror Deal Delayed a Gravitational Wave Detector

By Renu Shah / Jun 12, 2026

A €2–3 million precision mirror for Virgo was left unfunded, delaying the detector's upgrade by 18 months. The story reveals how rigid procurement rules and underbudgeted contingency can stall billion-euro science infrastructure.

One Grant Agency’s Animal-Derived Antibody Ban Complicates a Neurodegeneration Replication

By Renu Shah / Jun 12, 2026

Wellcome Trust’s 2025 ban on animal-derived antibodies disrupts a key Alzheimer’s replication study, raising questions about reproducibility gains versus reagent availability.

One Unversioned Random Seed Collapsed a Computational Sociology Agent-Based Model

By Jonas Eriksen / Jun 12, 2026

A single unversioned random seed caused an agent-based model of opinion dynamics to produce irreproducible results. Three replication attempts failed, sparking debate over seed reporting standards in computational science.

One Unreported Anesthesia Protocol Slowed a Whole-Brain Calcium Imaging Atlas

By Jonas Eriksen / Jun 12, 2026

A hidden confound in anesthesia protocols stalled a whole-brain calcium imaging atlas for nearly a year. The fix reveals how critical methodology is for large-scale neuroscience.

One Unrecorded Polymer Batch Number Skewed a Battery Cycling Study

By Jonas Eriksen / Jun 12, 2026

A missing lot number for a polymer binder skewed battery cycling data across labs for two years. The hidden variable cost US$400k and a retraction before anyone noticed.

One Untracked Awake-Asleep Transition Artifact Drove a Hippocampal Replay Finding

By Karim Osman / Jun 12, 2026

A 2006 hippocampal replay finding, long cited as evidence for memory consolidation, failed to replicate. Reanalysis reveals a subtle artifact from untracked awake-to-sleep transitions in spike sorting.

One Untracked Deep-Sea Thermistor Drift Bent a Decadal Ocean Heating Curve

By Jonas Eriksen / Jun 12, 2026

A single drifting thermistor on a deep Argo float skewed global ocean heat content estimates by 0.05°C over 15 years. A 2024 study corrects the record, reducing the apparent warming rate by 12% and tightening climate sensitivity constraints.

One Untracked Lab Diet Nutrient Shift Skewed a Mouse Behavior Battery

By Renu Shah / Jun 12, 2026

A choline-free chow switch in 2015 quietly altered mouse behavior baselines, exposing how untracked diet shifts can undermine reproducibility in behavioral neuroscience.

A Single Untracked Electrode Impedance Drift Inflated a Neural Recording's Yield

By Renu Shah / Jun 12, 2026

A 30% spike in neural yield traced to a loose connector reveals how untracked electrode impedance drift inflates unit counts, prompting a low-cost fix using voltage noise.

An Unfunded Database Maintenance Fee Fractured a Genomics Meta-Analysis

By Jonas Eriksen / Jun 12, 2026

A sudden access fee for genomic databases halted replication of 47 GWAS studies, shifting effect sizes and destabilizing cross-disciplinary research. The case exposes fragility in data commons funding.

One Unfunded Calibration Lab Closure Biased a Neural Recording Consortium

By Alice Chen / Jun 12, 2026

The closure of a national calibration lab introduced systematic bias into a multi-site neural recording consortium, undermining years of data on hippocampal replay.

One Untracked Social Desirability Screener Inflated a Morality Priming Replication

By Karim Osman / Jun 12, 2026

A single untracked social desirability screener added to a replication attempt of a morality priming study inflated an effect, sparking debate on methodological transparency.

One Uncalibrated Photometer Zero-Point Shift Silenced a Cepheid Distance Ladder

By Alice Chen / Jun 12, 2026

A tiny zero-point shift in a 1990s photometer introduced a systematic error that propagated through the Cepheid distance ladder, contributing to the Hubble constant tension.

One Grant Agency’s Scan-Time Cap Skewed a Whole-Brain Connectivity Atlas

By Alice Chen / Jun 12, 2026

A 12-minute scan-time cap imposed by a major grant agency inadvertently biased a widely used mouse brain connectivity atlas, leading to systematic undercounting of long-range neural projections.

An Unreported Stirring Rate Shift Doubled a Catalysis Lab’s Turnover Number

By Karim Osman / Jun 12, 2026

How a missed mixing parameter doubled catalytic yields, why labs ignored it for decades, and what a cheap protocol change means for chemistry reproducibility.

One Unrecorded Seawater pH Electrode Drift Masked a Pacific Acidification Pattern

By Alice Chen / Jun 12, 2026

A 0.02–0.03 pH unit drift in uncalibrated SeaFET electrodes masked a Pacific acidification trend. Jessica Cross's team corrected the data using a method borrowed from paleoceanography.

One Untracked Sediment Core Storage Fee Fractured a Paleoclimate Reanalysis Consortium

By Alice Chen / Jun 12, 2026

An unpaid $87 storage fee for a single sediment core box triggered the collapse of a major paleoclimate reanalysis consortium, highlighting the fragility of scientific infrastructure.

One Unrecorded Electrolyte Purity Lot Mismatch Inflated a Battery Paper’s Cycle Life

By Alice Chen / Jun 12, 2026

A trace impurity in one electrolyte lot doubled a battery paper's cycle life claims. The story of how a 0.1% mismatch led to retraction, and what it reveals about research incentives.

An Unversioned Solver Parameter Shift Reversed a Verified Climate Model Run

By Jonas Eriksen / Jun 12, 2026

A single solver tolerance change from 1e-8 to 1e-10 in a CESM library caused a 0.3°C temperature shift, unraveling a decade-old simulation. The 2019 audit by Baker et al. exposed how unversioned parameters threaten reproducibility in climate modeling.