Skip to main content
Tessa NewsLink
Paste a health news link, or browse

Source study found

Story checked

Experiment Tested Out-of-Body Experiences. Two Reported Seeing Things Scientists Can't Explain. : ScienceAlert (opens in a new tab)

Experiment Tested Out-of-Body Experiences. Two Reported Seeing Things Scientists Can't Explain. · 2026-09-11

Short answerEvidenceSource

Short answer

Mostly not supported

Mostly not supported.

One claim goes further than the study. 4 other points were not covered by the paper.

  • 1 supported
  • 1 overstated
  • 4 not covered

Checked against the study summary. The full text wasn't available, so some details couldn't be settled either way.

Share this check

Follow the evidence trail
1

The story

Experiment Tested Out-of-Body Experiences. Two Reported Seeing Things Scientists Can't Explain. : ScienceAlert

Experiment Tested Out-of-Body Experiences. Two Reported Seeing Things Scientists Can't Explain. · 2026-09-11

The story’s checkable claims.

Read the original story (opens in a new tab)
2

NewsLink checks it

Mostly not supported

One claim overstates the study. One of six checks out. Four claims the study doesn't address.

  • 1 supported
  • 1 overstated
  • 4 not covered
Open claim evidence
3
Source paper

Source layer

The 2 papers the story cites

Source study separated from background citations.

The research anchor for the report.

Then inspect each claim

Evidence layer

Claim by claim

Each claim gets a verdict. Expand it to see the evidence directly below.

6 claims in this story

Showing all 6 claimsChoose a verdict to focus the list.

Claim 1 of 6OverstatedTwo participants reported seeing something else that the researchers say they can't quite explain.View evidenceHide evidence

As statedtwo participants

Why this verdict

The profile supports that two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room. But the headline-style framing that they saw something 'impossible' that researchers 'can't quite explain' outruns the abstract profile, which frames the reports as anecdotal, methodologically informative, and warranting systematic follow-up rather than as unexplained anomalous perception. The story body caveats partly correct this, but the lead is stronger than the paper evidence.

1

Study evidence

Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.

“Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.”
Claim 2 of 6Not coveredA team led by neuroscientist Marina Weiler of the University of Virginia tested out-of-body experiences in a carefully blinded experiment with 21 experienced practitioners and a hidden image in another room.View evidenceHide evidence

As stated21 experienced out-of-body practitioners

Why this verdict

The abstract-level profile supports the core experimental setup: 21 OBE-institute-affiliated participants attempted to perceive a randomly selected concealed visual target in a separate inaccessible room, with targets unknown to participants and researchers. However, the profile does not verify the story's author/institution details, that Marina Weiler led the team, or the stronger wording that all 21 were 'experienced practitioners.'

1

Study evidence

No individual trial showed statistically significant above-chance correspondence between participant descriptions and the concealed visual target.

“Twenty-one individuals affiliated with an institute specializing in OBE training attempted to perceive a randomly selected target displayed in a separate, inaccessible room.”
Claim 3 of 6Not coveredNot a single participant succeeded at identifying the hidden target image.View evidenceHide evidence

As stated0 of 21 participants

Why this verdict

The profile supports a null result for target correspondence: no individual trial and no combined rank-based test showed above-chance performance. But the abstract-level evidence does not verify the absolute story claim that 0 of 21 participants 'succeeded' or correctly identified the target, and it notes that only 13 trials produced usable target descriptions for analysis.

1

Study evidence

No individual trial showed statistically significant above-chance correspondence between participant descriptions and the concealed visual target.

“Twenty-one individuals affiliated with an institute specializing in OBE training attempted to perceive a randomly selected target displayed in a separate, inaccessible room.”
Claim 4 of 6Not coveredAfter AI compared participant descriptions with all 100 possible target images, the correct image ranked from 16th to 92nd, which the article says was no better than guessing.View evidenceHide evidence

As stated100 possible target images; ranks 16th to 92nd

Why this verdict

The profile supports the use of embeddings/cosine similarity against 100 candidate targets and supports the conclusion that the analyzed descriptions did not show above-chance correspondence. However, the exact rank range stated by the story, 16th to 92nd, is not present in the abstract-level profile and cannot be verified at this evidence depth.

1

Study evidence

No individual trial showed statistically significant above-chance correspondence between participant descriptions and the concealed visual target.

“Twenty-one individuals affiliated with an institute specializing in OBE training attempted to perceive a randomly selected target displayed in a separate, inaccessible room.”
Claim 5 of 6Not coveredThe findings have been published in Explore.View evidenceHide evidence

Why this verdict

The supplied paper profile does not include journal-publication metadata confirming that the study was published in Explore, so this claim cannot be verified from the abstract-level profile.

Claim 6 of 6SupportedTwo participants gave spontaneous descriptions of what researchers were doing in Room 2, and those descriptions matched the actual session arrangements, but the researchers describe these observations as anecdotal and inconclusive.View evidenceHide evidence

As statedtwo participants

Why this verdict

The profile supports that two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room, and that these reports were incidental/anecdotal and warrant systematic investigation rather than serving as controlled evidence. The specific label 'Room 2' is not in the profile, but the substantive claim is aligned with the abstract evidence.

1

Study evidence

Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.

“Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.”
Then look for missing context

Context layer

What the story left out

Important study details the story did not include.

  • Only 13 trials produced usable target descriptions for the primary analyses.

    This is a material limitation because it qualifies the apparent 21-participant sample for the main target-identification analysis. The story presentation emphasizes 21 participants but does not mention that only 13 usable target-description trials were analyzed.

    From concealed-target attempted-OBE experiment with computational semantic similarity

  • Positive-control analyses indicated that the computational pipeline could identify correct targets when semantic correspondence was present.

    The story presentation does not mention the analytic validation/positive-control component, which is a secondary contribution in the profile.

    From in_silico positive-control analyses

  • Positive controls do not establish the pipeline's sensitivity to naturalistic human language.

    This limitation is material to interpreting the null AI/ranking result. The story says the AI system found no better-than-guessing performance, but the caveats listed do not acknowledge that the controls may not prove sensitivity to naturalistic participant descriptions.

    From in_silico positive-control analyses

  • The sample was not randomly drawn from the general population; participants were affiliated with an OBE-training institute.

    The story notes that participants were experienced OBE practitioners, but it does not present the sampling limitation that they came from an OBE-training institute and therefore may not generalize.

    From concealed-target attempted-OBE experiment with computational semantic similarity

6 things the story did carry across
  • Primary concealed-target OBE experiment with randomized target unknown to participants and researchers in a separate inaccessible room.
  • Twenty-one participants affiliated with an institute specializing in OBE training attempted the task.
  • Computational semantic-similarity pipeline compared participant descriptions with 100 candidate targets using embeddings, cosine similarity, and rank-based inference.
  • Primary target-identification results showed no individual statistically significant above-chance correspondence and no evidence of above-chance performance in the combined rank-based test.
  • Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.
  • The spontaneous adjacent-room reports were anecdotal, incidental, not prespecified controlled endpoints, and warrant systematic follow-up rather than supporting a firm anomalous-cognition conclusion.
Then read the study layer

Study layer

Study at a glance

Scan the study first. Expand only the parts you want to inspect.

Pieces of work

3

Evidence read

study summary

Lead result

human in vivo

1Lead resulthuman in vivoTest whether participants’ post-session descriptions after attempted out-of-body experiences (OBEs) match concealed visual targets beyond chance using a preregisterable-style computational semantic similarity/ranking pipeline.concealed-target attempted-OBE experiment with computational semantic similarityExpand

In plain English

Prospective concealed-target experiment testing whether free-text descriptions after attempted out-of-body experiences (OBEs) correspond to randomized visual targets beyond chance using pretrained language-model embeddings and rank-based within-trial comparisons. Twenty-one participants attempted to perceive targets displayed in a separate inaccessible room; 13 trials yielded usable descriptions. No individual trial or the pooled rank-based test across trials showed evidence of above-chance correspondence. Two participants spontaneously reported accurate details about experimenters in an adjacent room.

Key findings

  • No individual trial showed statistically significant above-chance correspondence between participant descriptions and the concealed visual target.
  • A combined rank-based test across all 13 trials showed no evidence of above-chance performance.
“Twenty-one individuals affiliated with an institute specializing in OBE training attempted to perceive a randomly selected target displayed in a separate, inaccessible room.”
What this piece can’t prove
  • Only 13 trials produced usable target descriptions for the primary analyses.
  • Participants were affiliated with an institute specializing in OBE training (sample not randomly drawn from a general population).
  • Control analyses do not establish the computational pipeline's sensitivity to naturalistic human language, limiting interpretation of null results.
2in silicoValidate/verify the computational semantic-similarity pipeline using control analyses where semantic correspondence is present (as a positive control for the analytic workflow).in silico positive-control analysesExpand

In plain English

The authors ran in‑silico positive-control analyses showing that the embedding + cosine similarity + rank-based workflow can recover correct targets when semantic correspondence is present; the abstract notes these controls validate the analytic pipeline but do not establish sensitivity to naturalistic human language and are distinct from the human-subject tests.

Key findings

  • Control analyses confirmed that the computational pipeline identified correct targets when semantic correspondence was present.
“Control analyses confirmed that the pipeline identified correct targets when semantic correspondence was present, though these controls do not establish sensitivity to naturalistic human language.”
What this piece can’t prove
  • Abstract lacks detail on the specific construction, number, and quantitative performance of the positive-control cases.
  • Unclear whether control cases capture the variability and ambiguity of naturalistic human descriptions.
  • Controls assess analytic validity rather than participant-level phenomena and thus cannot substitute for sensitivity analyses on human-generated text.
3human in vivoDescribe spontaneous reports of accurate details about experimenters’ locations/activities in an adjacent room during/around attempted OBEs as potentially methodologically informative observations.case reportExpand

In plain English

Within an attempted-OBE study, two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room; the authors note these anecdotal environmental reports may have methodological value and warrant systematic investigation.

Key findings

  • Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.
“Two participants spontaneously reported accurate details about experimenters' locations and activities in an adjacent room.”
What this piece can’t prove
  • Observation is anecdotal and arises from a small number of participants (N=2), limiting generalizability.
  • Reports were incidental and not obtained via a preplanned, standardized assessment or blinded verification procedure.
  • Abstract provides no details on timing, content specificity, or methods of corroboration for these spontaneous reports.

1 further detail could not be confirmed from the summary.

Finally, the search trail

Method layer

NewsLink found the paper. Tessa takes you deeper.

NewsLink checks the story. Tessa is where you inspect the paper, authors, evidence, and research context.

Papers considered

The selected paper, plus nearby candidates.

PubMed, Crossref, Europe PMC · 15 candidate papers

Candidate

A quasi-experimental comparison of outcome-based hybrid learning and traditional instruction in environmental health education for preventive medicine undergraduates from Guilin Medical University.

Frontiers in Public Health · 2026 · PubMed

Tessa (opens in a new tab)Source (opens in a new tab)

And 9 more candidates considered.