← home
simulation · 061

blindsight

what forced-choice reveals about processing without awareness

Patient DB had a scotoma — a cortical blind region in his left visual field, caused by occipital surgery. Standard perimetry confirmed it: he reported no sensation when stimuli were placed there. The finding was unremarkable. Cortical lesions produce blind regions.

Lawrence Weiskrantz ran a different test. He presented stimuli in the scotoma and asked DB to guess anyway — to give a direction response even when he reported nothing. DB found the request perplexing. He complied.

86 of 100 trials in the scotoma: correct. DB said he had seen nothing on 99 of them. Both statements were true.

the experiment

Each trial shows a brief stimulus somewhere in the patient's visual field. Two responses are recorded independently: a verbal report ("did you see anything?") and a forced-choice direction guess ("left or right?"). The scotoma region uses a subcortical pathway — superior colliculus → pulvinar → extrastriate cortex — that processes the stimulus below the threshold of awareness.

verbal report
forced-choice direction
waiting to start
speed:
performance

Statistics accumulate across trials. Each bar's right edge shows current percentage. The dashed line marks chance (50%) — the expected baseline for random guessing on a two-alternative forced-choice task.

normal field — 0 trials
reported seeing
50%
direction correct
scotoma — 0 trials
reported seeing
50%
direction correct
pattern emerging
what this shows

The two output channels — verbal report and motor direction response — come from different processing streams. The report tracks conscious experience accurately. The direction response tracks whether the stimulus was processed, regardless of whether it reached awareness.

DB's introspective report was not unreliable. It was accurate within its domain. The domain did not cover everything the visual system was doing. He said he saw nothing in the scotoma, and that was true. He also responded correctly 86% of the time, and that was also true. The gap between them generated no signal from inside.

The subcortical pathway — superior colliculus to pulvinar to extrastriate cortex — is an ancient visual processing route. It is tuned for motion and coarse spatial location rather than object recognition. It is not connected to the recurrent cortical amplification that generates perceptual awareness. It processes; it does not report.

DB's actual performance (Weiskrantz et al., 1974): 86/100 correct in the scotoma on forced-choice location tasks. He reported seeing the stimulus on exactly one trial. This simulation approximates those proportions. The normal field performance (90% direction, 92% reported) and scotoma performance (74% direction, 5% reported) are modeled, not exact.
see also: entry-529 — Guess Anyway