Pedro Lago

PID control strategies for the automatic control of neuromuscular blockade

Abstract A digital PID controller incorporating several modifications to accommodate the characteristics of the neuromuscular blockade is described in this paper. The initial design was performed by the Ziegler-Nichols step response method. The controller gains are automatically adjusted to the specified target by a gain-scheduling technique. The control system was evaluated in 30 patients. Subsequent analysis of all the data collected during surgery indicates that the variability of the patient responses is much wider than that inferred from the literature. This observation clearly suggests the desirability of individual tuning of the controller parameters. A technique for an improved tuning of the PID controller parameters to the patient’s individual dynamics is presented.

Circadian heart rate variability rhythm in shift workers

The objective of this study was to assess the influence of day-night cycle and sleep-awake period on the circadian pattern of heart rate variability (HRV). Twelve male oil refinery security shift workers, aged 39 +/- 7 years, were studied with 24-hour Holter monitor recordings during morning and night work periods. Hourly HRV parameters in the time and frequency domains were evaluated. For both shifts, all HRV parameters during awake or work periods were found not to be statistically different. In both day and night work shifts, the very low frequency and high-frequency components of HRV and the proportion of differences in successive R-R intervals greater than 50 ms increased during the sleep period, while the low frequency/high frequency ratio decreased. The low-frequency component in absolute units and the SD of the R-R interval did not show any variation in either shifts for the different periods. These results suggest that the circadian pattern of HRV seems to be predominantly related to sleep (supine) and wakefulness (standing) and remains independent of night-day cycle.

Heart rate variability in brain death

The sensitivity and specificity of heart rate variability (HRV) in the corroboration of brain death diagnosis in patients with acute traumatic intracranial lesions was evaluated in 20 patients with clinical criteria of brain death, nine patients in deep coma (Glasgow scale <7) and 18 normal controls, all age matched. The electrocardiogram was sampled at 650 Hz and several parameters of HRV were calculated, in both time and frequency domains. The HRV parameters were significantly lower in the brain death group compared with the deep coma group. Linear discriminant analysis between brain death and deep coma patients was performed on a data set made of nine randomly selected patients with clinical criteria of brain death and nine patients in deep coma. Cross-validation was performed on the remaining 11 patients with clinical criteria of brain death. All patients in the data set were correctly classified (sensitivity and specificity of 100%). All patients in the cross-validation set were correctly classified (sensitivity of 100%). Further studies are necessary to evaluate the specificity of the method in the independent set of deep coma patients and in the follow-up of comatose and vegetative patients to identify irreversibility of HRV. Nevertheless, these results suggest that HRV analysis constitutes a fully sensitive and specific method for assessing brain death in potential organ donors with acute traumatic lesions of the brain. This fast, quantitative and bedside method seems very promising for the early confirmation of brain death, which is an important factor for the success of transplantation procedures and could have a high predictive value of brain death in comatose patients with brain injuries without fully diagnostic criteria.

Hipocrates: a robust system for the control of neuromuscular blockade.

Objective. Development of an automatic system (software package Hipocrates) for the control of neuromuscular blockade by continuous infusion of the non-depolarising types of muscle relaxant drugs presently used in anaesthesia, namely atracurium, cisatracurium, vecuronium and rocuronium. Methods. Hipocrates incorporates control strategies based upon classical, adaptive and robust control, as well as a wide range of noise reduction techniques and on-line adaptation to patient-specific characteristics. Therefore, the system provides strong robustness to inter- and intra-individual variability of the patients responses or unexpected circumstances and adaptation to the individual requirements. Results. The control system is easy to set up and to use in a clinical environment. It consists of a portable PC computer, a Datex AS/3 NMT sensor and a B|Braun compact perfusion pump. In the simulation mode the software package incorporates sophisticated generation of pharmacokinetic/pharmacodynamic models driven by simulated drug administration regimes (bolus, continuous infusion and a combination of both). Conclusions. Hipocrates is an advanced stand-alone application for the control of neuromuscular blockade with a friendly graphic interface. It has been extensively validated, and it can be used on patients undergoing surgery as well as for simulation studies. Therefore Hipocrates also provides an excellent environment for education and training purposes.

Time domain baroreflex sensitivity assessment by joint analysis of spontaneous SBP and RR series

Abstract The sequences technique is frequently used for time domain assessment of the arterial-cardiac baroreceptor reflex sensitivity (BRS). The BRS is estimated by the slope between systolic blood pressure and RR interval values in baroreflex sequences (BSs) and an overall estimate is obtained by slope averaging. However, only 25% of all beats are in BSs with 60% of those located in 3-beat length segments. Also, in cases of BSs absence (usually associated with poor BRS function), the BRS cannot be quantified. Here, baroreflex events (BEs) are introduced and used with global/total slope estimators to improve BRS assessment. The performance of the novel method is evaluated using the EuroBaVar dataset. The events technique benefits from a higher number of beats: 50% of all beats are in BEs with more than 70% exceeding 3-beat length. It always provides a BRS estimate, even when BSs cannot be identified. When BSs are available, estimates from BEs and BSs are highly correlated. The estimates from BEs for the cases without BSs are lower than the estimates for the remaining cases, indicating poorer BRS function. The events technique also offers superior ability to discriminate lying from standing position in the EuroBaVar dataset (23/23 versus 18/23 for the sequences technique).

Heart rate variability in 24-hour holter recordings: Comparative study between short- and long-term time- and frequency-domain analyses

Mean hourly parameters obtained from all beats (long series) were compared with those obtained from a sample of 512 beats extracted each hour (short series) in nine presumably normal subjects. For both the short and long series, the spectral components, very low frequency, (VLF), low frequency (LF), and high frequency (HF), and time-domain indices (such as the Ewing statistic [PNN50] and RR standard deviation [SD-RR]), have been estimated. The spectral components LF and HF, estimated from the short and long series, were not significantly different, whereas significant differences were found between VLF, SD--RR, and PNN50. In both the short and long series, a strong correlation was found between LF and SD-RR and between HF and PNN50. The results suggest that, over a period of 24 hours, hourly LF and HF spectral components can be obtained using a single series of 512 beats every hour, with a great advantage over the evaluation of the mean hourly parameters. This method would be particularly useful in the study of circadian heart rate spectral analysis in Holter recordings with multiple artifacts or ectopic beats, and in general, when analysis of the entire 24-hour series is not feasible.

Comparison of On-line Autocalibration Techniques of a Controller of Neuromuscular Blockade

Abstract A method for the on-line autocalibration of a controller of muscle relaxation is presented. It relies on the information about the patients dynamics deduced from the response to the bolus of relaxant given in the beginning of anaesthesia. For this purpose, the bolus response is characterised by parameters such as the elapsed time between the bolus administration and the time the relaxation reaches 50% and the slope of the descent of the curve towards zero. A more complete description of the bolus response can be achieved by principle component analysis (PCA). A detailed comparison of these alternative descriptions is presented.

Control of Individualized Drug Dosage Regimens Based on Sparse Measurements

Abstract A number of different methods can be used for the design of individualized drug dosage regimens, based on measurements of the plasma concentration. This paper extends an adaptive control method coupled with Bayesian feedback to include all common routes and restrictions on the dosage regimens. Bolus administration and a piecewise constant administration (a model for the intravenous infusion) are both considered for pharmacokinetic systems of order n . Besides, a MATLAB package for the design of individualized drug dosage regimens was developed, both for clinical use and simulation purposes. The techniques are illustrated with the dosage of theophylline

Covariance density estimation for autoregressive spectral modelling of point processes

The use of autoregressive modelling has acquired great importance in time series analysis and in principle it may also be applicable in the spectral analysis of point processes with similar advantages over the nonparametric approach. Most of the methods used for autoregressive spectral analysis require positive semidefinite estimates for the covariance function, while current methods for the estimation of the covariance density function of a point process given a realization over the interval [0,T] do not guarantee a positive semidefinite estimate. This paper discusses methods for the estimation of the covariance density and conditional intensity function of point processes and present alternative computational efficient estimation algorithms leading always to positive semidefinite estimates, therefore adequate for autoregressive spectral analysis. Autoregressive spectral modelling of point processes from Yule-Walker type equations and Levinson recursion combined with the minimum AIC or CAT principle is illustrated with neurobiological data.

Dynamical Models of Muscles Using Emg-Force Data

Abstract The general field of muscle modelling is considered. Both non-linear and linear model structures are examined and their advantages and disadvantages discussed. A class of linear models based on EMG-force data derived in isometric conditions is discussed in detail. It is shown that attention has to be paid not only to the physical conditions of the test but also to the model structure, the estimation algorithm and test manoeuvre chosen as well as the sampling rate and the model conversion formula used. The basically non-linear, time varying nature of the system and the non-additive nature of the noise can all be accounted for.