Embedded Files
Our device consists of four main components: (1) Grip (2) Cable (3) Filter (4) Data Acquisition. The components come together as seen in the diagrams below.
(1) Grip
The grip is strapped to the patient's hand with an adjustable Velcro strap. The FlexiForce sensor is secured using double-sided tape to the surface of each of the pucks (in blue on the figure below on the left) and the screws are loosely fastened to help keep the motion limited to the vertical direction. The positioning of the sensor can be seen in our first prototype (the green one) below in the middle. The base of the grip rests on the side of the index finger (as seen in the image below on the right) and the patient can press with his or her thumb onto the top of the grip. The pucks ensure that the force is transmitted to the sensing area of the FlexiForce.
(2) Cable
The cable we chose is a shielded, twisted pair cable. In order to reach from the patient's hand to the penetration panel, the length needed to be about 25 feet. The cable is plugged into one of the penetration panel's D-Sub connectors, and on the other side of the panel, in the control room, the D-Sub connector is connected to the filter.
We tested both coax and twisted pair cable, and while they performed similarly outside the MRI, the twisted pair cable proved to shield from noise twice as well, as can be seen in the plot on the left, below. The data in the plot is baseline noise in each cable, post filtering, while the MRI was actively scanning. Additionally, it was determined that the twisted pair cable combined with the filter reduced noise 32 times compared to the baseline noise as seen in the plot below on the right.
(3) Filter
The filter used is a single-source, first order, active, low pass filter. We tested dual source, first order, active low pass filters first, but found that a single source worked just as well and didn't have the added complexity of needing two power sources. The filter is designed to have a cutoff frequency of 10 Hz to account for the 40 Hz sampling rate. After testing the filter on a breadboard, a PCB was designed for the final device (see below, left)
(4) Data Acquisition
The filtered signal is passed into a 16-bit National Instruments data acquisition system which is connected to a computer with LabView. LabView code developed by Dr. David Dubowitz of CFMRI, and modified by our team, reads the signal, outputs a plot of the voltage v. time, and saves a text file of data. The team wrote a MATLAB script that takes the text file, extracts the relevant information, plots a voltage v. time curve, uses the calibration equation to convert voltage to force, and then plots a force v. time curve.
Page updated
Report abuse