Why do better face recognizers use the left eye more?
Bibliographic record
Abstract
Blais et al. 2013 showed that the best participants in a facial emotion recognition task used the left eye of face stimuli more than the other participants. By inducing the use of the left or the right eye in different subjects, Gosselin et al. 2014 demonstrated that left-eye usage caused better face recognition. We hypothesized that this effect may result from the right hemisphere face processing superiority (e.g. Voyer et al. 2012). In Experiment 1, we replicated Gosselin et al. (2014) using a different induction method and a more controlled setting. Specifically, we induced the use of the left (N=15) or the right eye (N=15) during a gender discrimination task by eliminating the gender-diagnostic information from the other eye. Group classification images revealed that the informative eye was the only region significantly used (p< .01, Cluster test). Performance, as indexed by the number of bubbles required to reach 75% of correct responses, was not different in the two subject groups before (p=.5) or after (p=.13) the induction but the left-eye group performed significantly better than the right-eye group (F(1,28)=6.38, p=.01) during the induction. In Experiment 2, we examined whether this left eye performance effect is related to the right hemisphere face processing superiority. Twenty subjects did the same face gender categorization task as in Exp.1 except that an eye-tracker (Eyelink II, 250Hz) was used to enforce fixation at the center of the screen and that the induced eye was presented 2.2 deg to the left, to the right or under the fixation cross. Results show, as in exp.1, more efficient face processing for left-eye than for right-eye subjects, but only when faces were presented to the left and under the fixation cross (F(1,113)=16.81,p< 0.001 and F(1,113)=5.75, p=0.01 respectively), corroborating the right hemisphere face processing superiority hypothesis. Meeting abstract presented at VSS 2016
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.009 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.007 | 0.002 |
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 source (direct Gemma or distilled Codex), 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".