Dan Barlow's Webpage | |||||||
Science Experiments |
|||||||
I've always been fact based. My most recent employment was with Kelly Solutions, assigned to the National Institute of Mental Health. My assignment was to support the researchers by maintaining existing experiment software and hardware, and creating new experiments to specification. I became familiar with real time stimulus presentation and response collection, instrument synchronization, and data extraction and statistics calculation. I was responsible for everything necessary for conducting the experiment, from site wiring through instruments and computers to data analysis and incident response. One constant concern was the accuracy of timing in response collection. I designed and built a test jig to compare various fiber optic and USB response pads to the "gold standard" parallel port wired switch. Even using an ExpressCard parallel port, the experiment software can detect responses at millisecond accuracy. Standard USB keyboards not so much; 30 milliseconds delay with a wide standard deviation. Well engineered response devices have internal clocks and deliver a timestamp with the button key code.
I was also asked to provide proof of concept sensors for new experiments, such as this hand force sensor. I was responsible for keeping all the legacy sensors working. Many needed to operate inside the MRI scanner at 3 Tesla magnetic fields, without picking up or causing interference. I did extensive programming of experiments in a Visual Studio based package called E-Prime. I also wrote data extractions and scripts to synchronize data files from multiple collecting instruments. I use a script environment caled Tool Control Language/Tool Kit. You might be familiar with it as the graphics interface for Python. I use a distribution that includes SQLite for statistics processing and report generation. In the correct security environment it can open an Excel spreadsheet and create sheets for raw data, analysis, and graphs with everything correctly placed automatically. That work all involves PII, and I don't have any copies of it; but you would benefit from my experience in creating those tools. |