For the temperature sensor, I used the National Semiconductor LM35 precision temperature sensor, which is nice because it does not require any explicit calibration, accurate to +/- 1/4 degree (centigrade), and is low-power. The precision temperature amplifier is instantiated with just a standard op amp, using precision (1%) resistors, since I am interested in getting precise values off the sensor.
The GSR sensor basically measure the skin conductance/resistance between two electrodes. The circuit basically instantiates a wheatstone bridge to find the unknown skin resistance value.
Here's a scan of the temperature and gsr circuits:
The EKG is the most tricky to get right, because of the very small potentials involved (on the order of mVs) as well as susceptibility to noise. For this I used the INA321 Burr-Brown instrumentation amplifier, in conjuction with a couple of other op-amps for filtering. The differential amplifier gets rid of the common mode noise, and an additional amplifer is used to hi-pass filter the signal to get rid of the dc bias. The resulting signal is then further amplified by another op-amp set up to as a low-pass filter to get the final signal for the hoarder.
The circuitry for the EKG took a very long time to get working for my final project. One of the problems why it took longer to get the EKG working was the design for the EKG was more complicated than the GSR/temperature sensors, and there was much more debug required to get the circuits working given that the surface mount parts were very difficult solder on correctly without bridges and other hard-to-see mistakes.
Here's a scan of the ekg circuit:
