Extended Data Fig. 7: Neuroprosthetic baroreflex implemented in rodents with SCI.
From: Neuroprosthetic baroreflex controls haemodynamics after spinal cord injury
Step 1(a): We tested whether we could stabilize haemodynamics using the neuroprosthetic baroreflex, operating in a closed loop, in animals with acute (n = 7; 12 h after injury) and chronic (n = 6; one month after injury) T3 spinal cord injury. We implemented the neuroprosthetic baroreflex within research-grade technology to achieve precise control over stimulation parameters. Step 2(b): Schematic of electronic dura mater electrode arrays targeting haemodynamic hotspots. Bar chart shows relative pressor responses (systolic blood pressure; n = 5, independent samples one-tailed t-test; t = 3.90; P = 0.006). Step 3(c): We found a linear relationship between stimulation amplitude and the pressor response to stimulation in both animals with acute (mixed model linear regression; R2 = 0.84, P < 2.2 × 10−16) and chronic (mixed model linear regression; R2 = 0.81, P < 1.0 × 10−15) spinal cord injury. Step 4(d): We completed a series of trials to test the ability of the neuroprosthetic baroreflex to stabilize haemodynamics in rats with acute spinal cord injury. The neuroprosthetic baroreflex was activated in a closed loop before the activation of the lower-body negative-pressure chamber. In trials where the stimulation was ON, we found a reduction in the target (baseline) error (paired samples one-tailed t-test; systolic blood pressure (SBP): t = −6.12, P = 5.50 × 10−4; mean arterial pressure (MAP); t = −6.08, P = 4.48 × 10−4; diastolic blood pressure (DBP); t = −5.85, P = 4.34 × 10−4) reduced time outside key thresholds (−10 mmHg; paired samples one-tailed t-test; SBP: t = −12.52, P = 9.94 × 10−6; MAP; t = −12.29, P = 8.83 × 10−6; DBP; t = −11.73, P = 1.15 × 10−5), a restoration of the nonlinear relationship between blood pressure and chamber pressure, and a concomitant reduction in the linear model coefficient (likelihood ratio test of nested models; all P < 0.001). These quantifications held for systolic blood pressure (top), diastolic blood pressure (middle), and mean arterial pressure (bottom). Step 5(e): We completed the exact same experiments on animals with chronic spinal cord injury and found similar results to those of the acutely injured rats. Specifically, in trials where the stimulation was ON, we found a reduction in the target (baseline) error (paired samples one-tailed t-test; SBP: t = −3.84, P = 0.006; MAP; t = −3.83, P = 0.006; DBP; t = −3.83, P = 0.006), reduced time outside key thresholds (−10 mmHg; paired samples one-tailed t-test; SBP: t = −4.37, P = 0.004; MAP; t = −4.43, P = 0.003; DBP; t = −4.21, P = 0.004), a restoration of the nonlinear relationship between blood pressure and chamber pressure, and a concomitant reduction in the linear model coefficient (likelihood ratio test of nested models; all P < 0.001). These quantifications held for systolic blood pressure (top), diastolic blood pressure (middle), and mean arterial pressure (bottom). Step 6(f): We found that in response to stimulation blood pressure rapidly reached the set-point, with convergence times of 0.76 s in the example case presented in Fig. 3, and 1.15 s (95% confidence interval: 0.36–2.5 s) across n = 13 animals in response to the negative-pressure chamber. In this case convergence was defined as stable within 2.5 mmHg. Step 7(g): The neuroprosthetic baroreflex, acting in closed loop, re-established natural frequency dynamics (increased wavelet power in the 0.4–1.0 Hz spectrogram) in both animals with acute (paired samples one-tailed t-test; t = 4.46; P = 0.002) and chronic SCI (paired samples one-tailed t-test; t = 3.37; P = 0.014). *P < 0.05; **P < 0.01; ***P < 0.001.
