Fig. 1: Dynamics of the new odor-delivery device.

A We verified using the PID response to linalool that the odor-delivery device was capable of delivering sharp and short odor pulses. B On the contrary, adding a 15 cm glass tube after the valve produced responses that were much less sharp, and short stimuli (up to 200 ms) evoked very little PID response or no response at all (we used pure linalool instead of 10% dilution to compensate for airflow mixing in the glass tube). C More volatile compounds produced sharper PID responses. D Shaded area indicates linalool stimulation. Approximately 2.8 s after the stimulus onset a plastic barrier was dropped between the PID and the odor-delivery device to prevent any odor molecules from the odor-delivery device from reaching the PID. The offset of the PID signal remained slow. E We dropped the barrier at different times after the stimulus onset. The longer the stimulus was, the slower the PID response offset. We observed the same pattern when we used our odor-delivery device to deliver stimuli of different durations (F). G–J We compared the value (averaged in a 20 ms window) of the PID at different times after the stimulus offset to its peak value. At 0.5 s after the stimulus termination, the sustained signal was the same regardless of whether the stimulus was terminated with the electrovalve or mid-odor delivery with a plastic barrier. This shows that most of the slow dynamics observed with the PID were due to the properties of the PID and not the odor-delivery device. The linalool concentration delivered was, therefore, likely to be sharper than measured by the PID. All PID responses in the figure were filtered with 49 Hz 2-pole Butterworth lowpass filter to remove noise.