Dataset Information

A nonparametric method for detecting fixations and saccades using cluster analysis: removing the need for arbitrary thresholds.

ABSTRACT: Eye tracking is an important component of many human and non-human primate behavioral experiments. As behavioral paradigms have become more complex, including unconstrained viewing of natural images, eye movements measured in these paradigms have become more variable and complex as well. Accordingly, the common practice of using acceleration, dispersion, or velocity thresholds to segment viewing behavior into periods of fixations and saccades may be insufficient.Here we propose a novel algorithm, called Cluster Fix, which uses k-means cluster analysis to take advantage of the qualitative differences between fixations and saccades. The algorithm finds natural divisions in 4 state space parameters-distance, velocity, acceleration, and angular velocity-to separate scan paths into periods of fixations and saccades. The number and size of clusters adjusts to the variability of individual scan paths.Cluster Fix can detect small saccades that were often indistinguishable from noisy fixations. Local analysis of fixations helped determine the transition times between fixations and saccades.Because Cluster Fix detects natural divisions in the data, predefined thresholds are not needed.A major advantage of Cluster Fix is the ability to precisely identify the beginning and end of saccades, which is essential for studying neural activity that is modulated by or time-locked to saccades. Our data suggest that Cluster Fix is more sensitive than threshold-based algorithms but comes at the cost of an increase in computational time.

SUBMITTER: Konig SD

PROVIDER: S-EPMC4091910 | biostudies-literature | 2014 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A nonparametric method for detecting fixations and saccades using cluster analysis: removing the need for arbitrary thresholds.

König Seth D SD Buffalo Elizabeth A EA

Journal of neuroscience methods 20140206

<h4>Background</h4>Eye tracking is an important component of many human and non-human primate behavioral experiments. As behavioral paradigms have become more complex, including unconstrained viewing of natural images, eye movements measured in these paradigms have become more variable and complex as well. Accordingly, the common practice of using acceleration, dispersion, or velocity thresholds to segment viewing behavior into periods of fixations and saccades may be insufficient.<h4>New method ...[more]

PMID: 24509130

Similar Datasets

Project description:Objectives: Video head impulse test (v-HIT) is a quick, non-invasive and relatively cheap test to evaluate vestibular function compared to the caloric test. The latter is, however, needed to decide on the optimal side to perform cochlear implantation to avoid the risk on inducing a bilateral vestibular areflexia. This study evaluates the effectiveness of using the v-HIT to select cochlear implant (CI) candidates who require subsequent caloric testing before implantation, in that way reducing costs and patient burden at the same time. Study Design: Retrospective study using clinical data from 83 adult CI-candidates, between 2015 and 2020 at the Leiden University Medical Center. Materials and Methods: We used the v-HIT mean gain, MinGain_LR, the gain asymmetry (GA) and a newly defined parameter, MGS (Minimal Gain & Saccades) as different models to detect the group of patients that would need the caloric test to decide on the ear of implantation. The continuous model MGS was defined as the MinGain_LR, except for the cases with normal gain (both sides ≥0.8) where no corrective saccades were present. In the latter case MGS was defined to be 1.0 (the ideal gain value). Results: The receiver operating characteristics curve showed a very good diagnostic accuracy with and area under the curve (AUC) of 0.81 for the model MGS. The v-HIT mean gain, the minimal gain and GA had a lower diagnostic capacity with an AUC of 0.70, 0.72, and 0.73, respectively. Using MGS, caloric testing could be avoided in 38 cases (a reduction of 46%), with a test sensitivity of 0.9 (i.e., missing 3 of 28 cases). Conclusions: The newly developed model MGS balances the sensitivity and specificity of the v-HIT better than the more commonly evaluated parameters such as mean gain, MinGain_LR and GA. Therefore, taking the presence of corrective saccades into account in the evaluation of the v-HIT gain can considerably reduce the proportion of CI-candidates requiring additional caloric testing.

Dataset Information

A nonparametric method for detecting fixations and saccades using cluster analysis: removing the need for arbitrary thresholds.

Publications

A nonparametric method for detecting fixations and saccades using cluster analysis: removing the need for arbitrary thresholds.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets