TECHNICAL FEATURES OF A CCD-VIDEO-CAMERA SYSTEM TO RECORD CARDIAC FLUORESCENCE DATA

A charge-coupled device (CCD) camera was used to acquire movies of tra nsmembrane activity from thin slices of sheep ventricular epicardial m uscle stained with a voltage-sensitive dye. Compared with photodiodes, CCDs have high spatial resolution, but low temporal resolution. Spati al resolution in our system ranged from 0.04 to 0.14 mm/pixel; the acq uisition rate was 60, 120, or 240 frames/sec. Propagating waves were r eadily visualized after subtraction of a background image, The optical signal had an amplitude of 1 to 6 gray levels, with signal-to-noise r atios between 1.5 and 4.4. Because CCD cameras integrate light over th e frame interval, moving objects, including propagating waves, are blu rred in the resulting movies. A computer model of such an integrating imaging system was developed to study the effects of blur, noise, filt ering, and quantization on the ability to measure conduction velocity and action potential duration (APD). The model indicated that blurring , filtering, and quantization do not affect the ability to localize wa ve fronts in the optical data (i.e., no systematic error in determinin g spatial position), but noise does increase the uncertainty of the me asurements. The model also showed that the low frame rates of the CCD camera introduced a systematic error in the calculation of APD: for cu toff levels >50%, the APD was erroneously long. Both noise and quantiz ation increased the uncertainty in the APD measurements. The optical m easures of conduction velocity were not significantly different from t hose measured simultaneously with microelectrodes. Optical APDs, howev er, were longer than the electrically recorded APDs. This APD error co uld be reduced by using the 50% cutoff level and the fastest frame rat e possible.