matlab中trapz函数源代码-saccadeCurvature2002:Ludwig&Gilchrist(2002)随附的...

Matlab code that accompanies Ludwig & Gilchrist (2002). Measuring saccade curvature: A curve fitting analysis. Behaviour Research Methods, Instruments, & Computers. I apologise that the link given in that paper is no longer active. I still get requests for this code, so I thought I'd make it available here. I’m surprised no one else has made open-source code available to do this kind of stuff (e.g. in R or Python). We published this in 2002 when I was much younger and certainly much less wiser. I am pretty sure that if I had to write these functions afresh, I would be able to do so much more elegantly and efficiently. So don't be surprised if you see some clunky code. However, it should all work and the computations are sufficiently easy for it not to take ages. The code is reasonably well documented. The require a Nx4 sample matrix as input. In my applications, the four columns were: block number; trial (or saccade) number; sample x-coordinate; sample y-coordinate. However, the first column will essentially be ignored. All the functions will do is monitor the second column for a change in order to select the samples that belong to the current movement of interest. Once the samples have been selected you can calculate the various curvature metrics. The main functions of interest (to most people) would be ComputeCurve and CurveMetrics. The former computes the curvature using the curve fitting approach we advocated almost 20 years ago. The CurveMetrics function fits a whole bunch of metrics. Note that the integration method (area under the curve) metric is particularly crude (and, really, just wrong). If you want to use this metric, I suggest you use something more up-to-date such as the 'trapz' function in Matlab.