[Eeglablist] [EXTERNAL] Re: get_chanlocs
Eriksen, Jeff :LGS Neurodiagnostics
JEriksen at LHS.ORG
Fri Dec 27 15:25:25 PST 2019
Very helpful information. A few follow-up questions:
1. I do not see the “original” scanner offered on the website, so I think you are suggesting I will need the Mark II, for which I will make sure to get the proper bracket. However, the Mark II says it only has the 160 deg mono, whereas the “Core” has 85-deg color AND 160-deg mono, as best I can tell from the website specs. You write about having color, so this is still not crystal clear to me. I think I will have to contact the manufacturer directly.
2. The Apple Scanner App’s limitation to sending its data by email is new to me. I did not notice that on the EEGlab Wiki. Can you please give more details on this process? At what stage in the scanning process does it send the email? It sounds like a RAM and/or disk storage limitation. Are multiple emails sent during the scanning process? Any possibility that an iPad with more RAM or a bigger disk drive would obviate the need to email?
From: Seyed Yahya Shirazi <shirazi at ieee.org>
Sent: Friday, December 27, 2019 2:02 PM
To: eeglablist at sccn.ucsd.edu
Cc: Eriksen, Jeff :LGS Neurodiagnostics <JEriksen at LHS.ORG>; Andrew Engell <engella at kenyon.edu>
Subject: [EXTERNAL] Re: get_chanlocs
External Sender – Be Suspicious of Attachments, Links, and Requests for Payment or Login Information.
Dear Jeff and Andrew,
Glad to hear there is more and more interest in 3D scanning and EEG digitization/positioning in general. We do digitize EEG electrode locations using Structure Sensor and iPad Pro 10.5" in our lab.
Our recent paper compared 3D scanning with Zebris and motion capture, and Structure performed very well both in digitization reliability and also consistency in finding the same source locations (see https://www.ncbi.nlm.nih.gov/pubmed/31787866<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ncbi.nlm.nih.gov%2Fpubmed%2F31787866&data=02%7C01%7CJEriksen%40lhs.org%7C187a103e2348419178fc08d78b18781c%7C3683988d7d404338bf200b5dd13f4301%7C0%7C0%7C637130809539475855&sdata=6wSPMvIoap0QhW3wBLvmcsLoSL31DIUUwiO5lcqGclg%3D&reserved=0>).
Electrode digitization with a 3D scanner has two main steps: 1- scanning the head 2- marking the electrodes and fiducial locations in a software (here MATLAB). I would say after learning how to use the scanner, each step takes about 10 minutes with a 128-channel BioSemi ActiveTwo system. Gaia Taberna (from Dr. Manitini's lab in KU Leuven) also created a toolbox for automatically marking the electrodes, and they had the EGI system (see: https://www.ncbi.nlm.nih.gov/pubmed/31492919<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ncbi.nlm.nih.gov%2Fpubmed%2F31492919&data=02%7C01%7CJEriksen%40lhs.org%7C187a103e2348419178fc08d78b18781c%7C3683988d7d404338bf200b5dd13f4301%7C0%7C0%7C637130809539475855&sdata=AxjkwlmjmAfFrffRIe0JGgAX%2BiQ02CWeLg6NN9Hc4%2FA%3D&reserved=0>). I have not tried the automatic approach yet because 10 minutes is not a long time to do mark up the scan.
The original (NOT the core) scanner works with my iPad Air A2152 (it has the same chassis as iPad Pro 10.5 A1701). I am not sure about the earlier iPads, but I think as long as the Structure.io website claims the scanner supports that model, you are good to go. However, make sure that you have the bracket (adapter) to attach the scanner to the iPad (I checked strucutre.io<https://nam01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fstrucutre.io%2F&data=02%7C01%7CJEriksen%40lhs.org%7C187a103e2348419178fc08d78b18781c%7C3683988d7d404338bf200b5dd13f4301%7C0%7C0%7C637130809539485812&sdata=GEOlNtC5QyIe35Ki1IzTm7IwABc54xlDeQHWJxibZOg%3D&reserved=0>, and it seems they don't offer a bracket for A1474). Structure scanner MarkII is just a better scanner with better resolution. The core version is supposedly the standalone scanner with a tripod, so I don't think that's the option you want.
You only need the Scanner app from Apple AppStore (https://apps.apple.com/us/app/scanner-structure-sdk/id891169722<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fapps.apple.com%2Fus%2Fapp%2Fscanner-structure-sdk%2Fid891169722&data=02%7C01%7CJEriksen%40lhs.org%7C187a103e2348419178fc08d78b18781c%7C3683988d7d404338bf200b5dd13f4301%7C0%7C0%7C637130809539485812&sdata=bOfgkAajDOGuf%2FzcPoMm6rlzSJ1cPNtlhhuiIeJNovE%3D&reserved=0>) to do the scanning. The app is free and would ask you to install its calibrator app (also free) if you are scanning for the first time. The only downside of this app is that it can't store the scan on the iPad and only would send it via email, so you need an internet connection. Both EEGLAB and FieldTrip toolbox can read outputs of the Scanner app after you unzip it from the email attachment (see our pipeline, based on FieldTrip, here: https://github.com/neuromechanist/eLocs<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fneuromechanist%2FeLocs&data=02%7C01%7CJEriksen%40lhs.org%7C187a103e2348419178fc08d78b18781c%7C3683988d7d404338bf200b5dd13f4301%7C0%7C0%7C637130809539495769&sdata=HEYK%2BxPHrqvPc8ocv7%2BynglRDlZAkaWdKAvcUFkjJRQ%3D&reserved=0>).
The Structure sensor has its limitations too. With the original model that we are currently using, the scanner is susceptible to the lighting conditions, so we only do the digitization in a specific spot in our lab to make sure that they are similar. Also, sometimes the scanned point cloud and the color image that should be overlayed to the point cloud do not match, or the color image is not clear. So we always check the results on the iPad to make sure that we have the right quality color image, and the image matches the point cloud. As I said earlier, using a 3D scanner for EEG electrode digitization has a learning curve, but once you get used to the process, it just takes 10 minutes to scan and another 10 minutes to mark up the locations in MATLAB.
Please do not hesitate to contact me if I can be of more assistance.
Seyed Yahya Shirazi
Ph.D. Candidate, BRaIN Lab
University of Central Florida
More information about the eeglablist