I have a doubt about Performance measurements in the MIT Manus as they are described in : "Rehabilitation robotics: Performance-based progressive robot-assisted therapy". Although I was able to reproduce the PM2 measurements of the figure 6, the chart for the PM3 is very unlikely to corresponed to the equation (10) of the text. All numbers in that formula are positive, but the chart has negative numbers. The only hint in the text (page 14 second column) is that "PM3 is defined solely as a funcition of PM3b". Does anybody know how this PM3 is built?

https://link.springer.com/article/10.1023/A:1024494031121