JaneliaSciComp / JRCLUST

JRCLUST
Apache License 2.0
34 stars 22 forks source link

unusual distribution of center sites #124

Open thomasslai opened 5 years ago

thomasslai commented 5 years ago

when splitting clusters, the mean waveform displayed is that of the lowest channel/site number in the cluster, when it should be the max site. this leads to some questionable splitting when the lowest channel is not picking up the full waveform

If you have a usage or support question, come join our Gitter and ask your question there. If you found a bug, please provide all of the following:

JRCLUST v4.0.0-beta-p1 "Edward"

Output of ver

ver

MATLAB Version: 9.3.0.713579 (R2017b) MATLAB License Number: 40523909 Operating System: Microsoft Windows 7 Enterprise Version 6.1 (Build 7601: Service Pack 1) Java Version: Java 1.8.0_121-b13 with Oracle Corporation Java HotSpot(TM) 64-Bit Server VM mixed mode

MATLAB Version 9.3 (R2017b) Simulink Version 9.0 (R2017b) Aerospace Blockset Version 3.20 (R2017b) Aerospace Toolbox Version 2.20 (R2017b) Antenna Toolbox Version 3.0 (R2017b) Audio System Toolbox Version 1.3 (R2017b) Bioinformatics Toolbox Version 4.9 (R2017b) Communications System Toolbox Version 6.5 (R2017b) Computer Vision System Toolbox Version 8.0 (R2017b) Control System Toolbox Version 10.3 (R2017b) Curve Fitting Toolbox Version 3.5.6 (R2017b) DSP System Toolbox Version 9.5 (R2017b) Data Acquisition Toolbox Version 3.12 (R2017b) Database Toolbox Version 8.0 (R2017b) Datafeed Toolbox Version 5.6 (R2017b) Econometrics Toolbox Version 4.1 (R2017b) Embedded Coder Version 6.13 (R2017b) Filter Design HDL Coder Version 3.1.2 (R2017b) Financial Instruments Toolbox Version 2.6 (R2017b) Financial Toolbox Version 5.10 (R2017b) Fixed-Point Designer Version 6.0 (R2017b) Fuzzy Logic Toolbox Version 2.3 (R2017b) Global Optimization Toolbox Version 3.4.3 (R2017b) HDL Coder Version 3.11 (R2017b) HDL Verifier Version 5.3 (R2017b) Image Acquisition Toolbox Version 5.3 (R2017b) Image Processing Toolbox Version 10.1 (R2017b) Instrument Control Toolbox Version 3.12 (R2017b) LTE System Toolbox Version 2.5 (R2017b) MATLAB Coder Version 3.4 (R2017b) MATLAB Compiler Version 6.5 (R2017b) MATLAB Compiler SDK Version 6.4 (R2017b) MATLAB Report Generator Version 5.3 (R2017b) Mapping Toolbox Version 4.5.1 (R2017b) Model Predictive Control Toolbox Version 6.0 (R2017b) Model-Based Calibration Toolbox Version 5.3 (R2017b) Neural Network Toolbox Version 11.0 (R2017b) OPC Toolbox Version 4.0.4 (R2017b) Optimization Toolbox Version 8.0 (R2017b) Parallel Computing Toolbox Version 6.11 (R2017b) Partial Differential Equation Toolbox Version 2.5 (R2017b) Phased Array System Toolbox Version 3.5 (R2017b) Polyspace Bug Finder Version 2.4 (R2017b) Polyspace Code Prover Version 9.8 (R2017b) Powertrain Blockset Version 1.2 (R2017b) RF Blockset Version 6.1 (R2017b) RF Toolbox Version 3.3 (R2017b) Risk Management Toolbox Version 1.2 (R2017b) Robotics System Toolbox Version 1.5 (R2017b) Robust Control Toolbox Version 6.4 (R2017b) Signal Processing Toolbox Version 7.5 (R2017b) SimBiology Version 5.7 (R2017b) SimEvents Version 5.3 (R2017b) Simscape Version 4.3 (R2017b) Simscape Driveline Version 2.13 (R2017b) Simscape Electronics Version 2.12 (R2017b) Simscape Fluids Version 2.3 (R2017b) Simscape Multibody Version 5.1 (R2017b) Simscape Power Systems Version 6.8 (R2017b) Simulink 3D Animation Version 7.8 (R2017b) Simulink Check Version 4.0 (R2017b) Simulink Code Inspector Version 3.1 (R2017b) Simulink Coder Version 8.13 (R2017b) Simulink Control Design Version 5.0 (R2017b) Simulink Coverage Version 4.0 (R2017b) Simulink Design Optimization Version 3.3 (R2017b) Simulink Design Verifier Version 3.4 (R2017b) Simulink Desktop Real-Time Version 5.5 (R2017b) Simulink PLC Coder Version 2.4 (R2017b) Simulink Real-Time Version 6.7 (R2017b) Simulink Report Generator Version 5.3 (R2017b) Simulink Requirements Version 1.0 (R2017b) Simulink Test Version 2.3 (R2017b) Spreadsheet Link Version 3.3.2 (R2017b) Stateflow Version 9.0 (R2017b) Statistics and Machine Learning Toolbox Version 11.2 (R2017b) Symbolic Math Toolbox Version 8.0 (R2017b) System Identification Toolbox Version 9.7 (R2017b) Tracking and Sensor Fusion Toolbox Version 1.0 (R2017b) Trading Toolbox Version 3.3 (R2017b) Vehicle Network Toolbox Version 3.4 (R2017b) Vision HDL Toolbox Version 1.5 (R2017b) WLAN System Toolbox Version 1.4 (R2017b) Wavelet Toolbox Version 4.19 (R2017b)

Output of gpuDevice

no GPU

The command you used

(e.g., jrc detect-sort myparam.prm) jrc manual

If manually sorting, what it is you tried to do

splitting

Console output

aliddell commented 5 years ago

Where do you get that impression? By my reading, the split dialog is showing the mean waveform on the current site in the curation GUI, which is the max site unless you've manually switched sites in the time view. The splitting itself, provided you use multisite (which is the default), is done using the traces on all sites in the unit's site neighborhood.

thomasslai commented 5 years ago

Thanks for the reply! The default site in the time view was site 1 here. I switched it to site 6 for demonstration purposes, but in the split dialog it still only displays site 1, or at least that's what it says there (see green circles).

jrc_split

aliddell commented 5 years ago

Thanks for the update. I see I passed the label for the wrong site, but it does indeed show the mean waveform on the currently-selected site. You can tell by looking at the mean waveform in the split figure, and seeing that it looks much more like the one on site 6 than the one on site 1.

The real issue is why we have site 1 as the primary for this cluster and not site 3. Can you do a histogram of spikeSites for spikes in this cluster?

thomasslai commented 5 years ago

I wasn't sure how to select spikeSites for a specific cluster, so here's the histogram of the full res.spikeSites. Apparently it only contains a subset of the available sites and it looks like these are those sites with the lowest sitenr. of the group of clusters, e.g. 1 for the first group, 9 for the second group, 17 for third etc (see screenshots).

hist_res_spikeSites

unique_spikeSites

screen

aliddell commented 5 years ago

That's pretty strange. Can I see the data/prm file?

thomasslai commented 5 years ago

Sorry for the delayed response. You can find the .dat and the .prm file under the following link:

https://www.dropbox.com/sh/sm3y3vjn6j4rc1l/AADVYjv0AaA9DxmT0y9knw3ya?dl=0

Thanks!