Time to wave goodbye to phase scrambling - creating unrecognizable control stimuli using a diffeomorphic transform.
Bibliographic record
Abstract
To isolate the neural events associated with the recognition of visual objects from the events corresponding to the perceptual processing prior to object recognition, control stimuli are needed that contain the same perceptual properties as the objects, but are not recognizable. Traditionally, control stimuli have been generated using phase-scrambling, box scrambling, and more recently, texture scrambling. We show these methods yield poor control stimuli because they dramatically change basic visual properties (e.g., spatial frequency, perceptual organization) to which even the earliest stages of visual processing are sensitive. To overcome this limitation, we applied a new warping method, using a diffeomorphic transformation that preserves lower-level perceptual properties while removing meaning; we acquired norms for recognition at various degrees of warping (N=415 participants completed 15600 trials). We hypothesized that images warped using our new method will produce neural activity at pre-recognition stages along the visual hierarchy that is similar to the intact versions of these images. To test this hypothesis, we computationally modeled neural activity (using the HMAX model) for each distortion method and compared it to the intact version of the image at three stages along the visual system: simple and complex V1 cells, anterior areas corresponding to area V4, and inferior temporal cortex. Based on average neural output and the distribution of activity across simulated neurons, we found that "diffeomorphed images" were markedly more similar to intact images than any of the other distortion methods. Our results show that unrecognizable diffeomorphed images better match the fundamental visual properties of intact images, and therefore serve as more appropriate control stimuli in neuroimaging research. We suggest that diffeomorphed images should be used to disentangle the representation of perceptual and semantic object features during perception, memory and attention. Meeting abstract presented at VSS 2013
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".