Effect of technical settings and semen handling upon motility characteristics of dog spermatozoa measured using computer-aided sperm analysis.

Technical aspects of computer-aided sperm analysis and the influence of semen preparation were investigated for their effect on the measured motility characteristics of dog spermatozoa. Altering the internal image settings significantly influenced the measured motility by changing the ability of the computer to recognize spermatozoa. Similarly, the use of a longer minimum track point (the minimum length of sperm track detected before analysis) resulted in failure to detect some of the faster moving spermatozoa. There was a clear interaction between the search radius (the threshold distance below which objects are linked together) and the minimum track point. A 1 min analysis period was required to eliminate reduced motility as a result of sample deterioration upon the microscope stage. The dilution of semen to between 1:10 and 1:20 was necessary to allow accurate detection of sperm motility; however, such dilution significantly altered the motility characteristics of spermatozoa. The influence of viscosity and ionic composition of the media was confirmed by comparing dilution in seminal plasma with dilution in iso-viscous methylcellulose and iso-osmotic saline, respectively. Analysis temperature had a significant influence on sperm motility, although values were most constant within the range of 25-45 degrees C. Extremes of temperature had marked deleterious effects. Careful selection of internal image settings, the minimal track point and search radius, and the analysis time are essential for accurate detection of sperm motility. Moreover, dilution of spermatozoa per se, and dilution with media of different viscosities and ionic compositions can alter the sperm motility. Once these aspects of computer image analysis are determined for each system, the method can achieve a high degree of repeatability with interanalysis coefficients of variation of < 12%, and intra-analysis coefficients of variation of < 3% for most parameters.