1. ¹King's College London, Department of Clinical Neurosciences (Neurosurgery), London, UK
2. ²Department of Bioengineering, Imperial College, London, UK
3. ³Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, Massachusetts, USA

Correspondence: Dr AJ Strong, King's College Hospital, Department of Neurosurgery, London SE5 9RS, UK. E-mail: Anthony.strong@kcl.ac.uk

Received 7 July 2005; Revised 11 September 2005; Accepted 19 September 2005; Published online 2 November 2005.

Abstract

Laser speckle imaging of the exposed cerebral cortex allows detailed examination of the time course and topography of perfusion under different experimental conditions. Here we examine the quantitative capacity of the method and its sensitivity for the detection of peri-infarct depolarisations (PIDs). In four cats anaesthetised with chloralose, the right hemisphere was exposed and the right middle cerebral artery was occluded. The brain was illuminated with a laser diode, the speckle pattern was imaged, and images of inverse speckle correlation time (ICT) were derived from the calculated speckle contrast images. We examined the relationship of ICT with perfusion, as imaged quantitively using umbelliferone clearance (CBF_umb). Values of ICT and CBF_umb were compared and regression parameters were calculated for each experiment. In eight cats, cortical surface direct current (DC) potential was monitored at two locations and detection of PIDs by DC potential and ICT change was compared. ICT- and CBF_umb-derived values of perfusion were closely correlated, with a high degree of significance (P<0.0001). Overall, monitoring of DC potential detected 90% of PIDs, whereas ICT detected 56%. We conclude that (1) laser speckle imaging provides an index of perfusion that has a linear relationship with the clearance rate of umbelliferone within the range of levels of perfusion examined; (2) this relationship is relatively stable between experiments; and (3) the method's ability to detect blood flow changes associated with PIDs likely depends on the noise level of the speckle measurements.