Correlations Between the EEGs of Two Spatially Separated Subjects β A Replication Study
β‘ Contested βπ Appears in:
Plain English Summary
Can one person's brain activity ripple across to someone in a separate room? Earlier studies by Wackermann claimed yes -- they put pairs in shielded rooms eight meters apart, had one watch flashing patterns while the other relaxed, and reported mysterious brainwave correlations. But when Ambach tried replicating with seventeen pairs, he found something more interesting than telepathy: a bug in the math. The original statistical method flagged false positives roughly 20-25% of the time when it should have caught them only 5%. Once he applied a corrected approach, every single test came back non-significant. The kicker? The old flawed method applied to his own data would have produced seemingly impressive results. A cautionary tale about how subtle analytical choices can manufacture evidence for brain-to-brain connections that aren't there.
Research Notes
An important methodological contribution to the EEG-correlation/DMILS literature: identifies a specific statistical flaw that inflated false positive rates in prior studies by Wackermann et al. (and potentially other groups using similar bootstrapping). The failed replication and statistical correction together cast doubt on the prior positive findings and highlight the sensitivity of these paradigms to analytic choices. Directly relevant to Controversy on brain-to-brain interaction.
Replication of Wackermann et al.'s (2003, 2004) EEG correlation paradigm at IGPP Freiburg with a different experimenter. Seventeen pairs of related subjects were tested in acoustically and electromagnetically shielded rooms 8 meters apart; one subject viewed checkerboard reversals while the other relaxed. Critical review of the original bootstrap sampling method revealed it systematically overestimates effects (false positive rate 20.1-26.2% at nominal alpha = 0.05). A corrected nonparametric method with Monte Carlo correction for inter-channel dependence yielded no significant results across any of six tests (all p > .067). The original uncorrected method applied to the same data would have produced apparently significant results (1000 ms window: p = .038 uncovered, p = .0003 difference).
Related Papers
Replication Of
Critiques
Companion
- Electroencephalographic Evidence of Correlated Event-Related Signals Between the Brains of Spatially and Sensory Isolated Human Subjects β Standish, Leanna J (2004)
- Event-Related Electroencephalographic Correlations Between Isolated Human Subjects β Radin, Dean I (2004)
- Searching for Neuronal Markers of Psi: A Summary of Three Studies Measuring Electrophysiology in Distant Participants β Hinterberger, T (2010)
More in Telepathy
Telecommunication Telepathy: A Meta-Analysis
Rethinking Communication and Consciousness: Lessons from The Telepathy Tapes Podcast
Who's Calling? Evaluating the Accuracy of Guessing Who Is on the Phone
A Comparison of Four New Automated Telephone Telepathy Tests
Detecting Telepathy: A Meta-Analysis for Extrasensory Perception Experiments in Last 20 Years
π Cite this paper
Ambach, Wolfgang (2008). Correlations Between the EEGs of Two Spatially Separated Subjects β A Replication Study. European Journal of Parapsychology.
@article{ambach_2008_correlations,
title = {Correlations Between the EEGs of Two Spatially Separated Subjects β A Replication Study},
author = {Ambach, Wolfgang},
year = {2008},
journal = {European Journal of Parapsychology},
}