VOLTAGE VS FORCE IN A DENTAL EXPERIMENT

The data set zmk.txt consists of data on n=150 measurements. Each measurement relates electrical voltage in the chewing muscle, M. masseter, with the corresponding force exercised by the muscle and measured as pressure on a probe on which subjects bite during the experiment. A sensor attached to the probe measures force exercised. Subjects were asked to adjust their biting pressure in such a way as to achieve a fixed prescribed pressure which was varied. Then the corresponding voltage in muscle tissue was recorded; the relation of this voltage to the force generated by the muscle was of interest. Voltage is the dependent variable and force the independent variable. The data were recorded in the dental clinic of the University of Marburg, Germany, around 1986.

The regression relation describing an electro-physiological characteristic of muscular contraction involves both non-linearity and marked heteroscedasticity. The analysis of these data with kernel smoothing was discussed in Müller (1988) and with a nonparametric quasi-likelihood method assuming a known functional relation in Chiou & Müller (1999). Earlier analyses involved issues of local bandwidth choice, of estimating a variance function, and of quasi-likelihood modelling.

References:

Chiou, J., Müller, H.G. (1999). Nonparametric quasi-likelihood. Ann. Statist. (in press).

Müller, H.G. (1988). Nonparametric regression analysis for longitudinal data. Lecture Notes in Statistics vol. 46, Springer, pp. 36-38.

Data format:

First column, force(pressure); second column, voltage.