Randomized double pulse stimulation for assessing stimulus frequency-dependent conduction in injured spinal and...

Injury compromises the ability of axons to conduct action potentials at high frequencies. To study stimulus frequency-dependent conduction in injured spinal and peripheral axons, we developed a new stimulation paradigm which applies trains of double pulses at 5 Hz and randomly varied interpulse intervals of 3, 4, 5, 8, 10, 30, 50, and 80 msec. In each double pulse, the first pulse was used to condition the response activated by the second test pulse. Responses elicited by double pulses with 80 msec intervals served as controls. The L5 dorsal root was stimulated to activate dorsal column and dorsal root compound action potentials in pentobarbital anesthetized rats. To injure the spinal cord, we compressed the cord stepwise (0.25 mm every 5 min) until action potential conduction across the compression site was abolished and then decompressed the spinal cord 10 min later. Before injury, conditioning pulses applied 3-80 msec before the test pulses did not alter dorsal column responses except for a slight amplitude augmentation at 20 msec interpulse intervals (mean +/- S. E., + 4.2 +/- 0.8%, P less than 0.02) compared to controls. Injury had 3 effects on the responses. First, it significantly reduced response amplitudes and increased response latencies at 3-5 msec interpulse intervals, i.e., responses activated with 3 msec intervals were 26.0 +/- 7.4% (P less than 0.002, paired t test, n = 6) smaller and had 108 +/- 45 microseconds (P less than 0.04) longer latency than control responses. Second, response amplitude increases at 20 msec interpulse intervals (9.0 +/- 0.7%, P less than 0.0001) significantly exceeded those observed before injury (P less than 0.02, paired t test). Third, injury accentuated response amplitude declines during the stimulus train, most prominently at 80 msec intervals. Spinal cord injury did not affect the dorsal root responses. L5 root compression injury depressed dorsal root action potentials at 3-5 msec interpulse intervals (36.9 +/- 8.4%, n = 4, P less than 0.0001) but had no other effect on the responses. Our data indicate that randomized double pulse evoked potentials are sensitive detectors of acute axonal dysfunction and can be used to quantify stimulus frequency-dependent conduction deficits in injured central and peripheral axons.