Manipulations of visual feedback during handwriting result in a lengthening of reaction time (RT) and movement time (MT) amongst several other reported effects (see for example, Smyth and Silvers, 1987 and Van Doorn and Keuss, 1992. In line with a chronometric analysis of performance, these findings have been explained following the traditional resource theory according to which increments in RT and MT indicate increased processing demands. However, Van Gemmert and Van Galen, 1998 and Van Gemmert and Van Galen, 1997 showed that, when processing demands are varied, biomechanical adaptation strategies of the motor system can be observed as well and, moreover, such biomechanical adaptations, like increasing limb stiffness, prevent an increase of chronometric measures within the normal variation of task difficulty. Furthermore, investigators of human movement and performance have recently argued that the traditional chronometric, computational approaches did not recognize the role of biomechanic parameters of movement. For example, studies in which on-line visual feedback was unexpectedly manipulated during movement execution showed fast, flexible reactions of the motor system, urging the need for a more dynamic view of the role between perception and action in handwriting performance.

In the present paper we will explore how increased processing demands are reflected in both the chronometric and the biomechanic aspects of handwriting. In our experiments, processing demands are manipulated by sudden on-line transformations of the visual feedback of writing slant or size. Van Gemmert and Van Galen (1994) proposed a theory that increased processing demands reduce signal-to-noise ratios (SNRs) at the neural level, resulting in an increase of neuromotor noise in movement velocity profiles. This is exemplified, amongst others, by a higher incidence of movement dysfluencies, i.e., fluctuations in the velocity profile. The theory assumes that the motor system can compensate for deteriorated SNRs by a biomechanical noise-filtering strategy of increased limb stiffness. Limb stiffness may be varied in different ways but one such strategy is to increase simultaneously contraction of both the agonist and antagonist muscles. Evidence in favor of this theory is that participants use a higher degree of limb stiffness in task situations which require additional processing demands, such as high accuracy demands. The increased tonic limb stiffness filters the high neuromotor noise levels in the motor system at the cost of movement speed. In writing tasks, this strategy results in an increase of axial pen force. In contrast, resource theories conventionally ignore the role of biomechanic aspects of movements, emphasizing that an increment in chronometric measures is the only indicator of increased processing demands.

In summary, our theory assumes that transformations of the visual feedback of writing slant or size will cause an increase of processing demands, which in turn causes deteriorated SNRs, resulting in increased dysfluencies in the motor output signal. To filter amplified levels of noise, the motor system increases the overall level of limb stiffness, which results in an increase of axial pen force at the cost of movement speed. At the same time, however, inherent to changes in size, the applied level of axial pen force may also change, because larger writing sizes are realized by higher force levels but not by longer writing times. Therefore, we first transformed the visual feedback of slant without transforming size. In this experiment the visual feedback of the writing slant was transformed by a forward or backward transformation relative to the participants’ normal writing slant. Second, we transformed the visual feedback of writing size without slant distortion, in order to unravel the differences between increments in axial pen force due to biomechanical noise-filtering and changes in axial pen force caused by different writing formats. In Experiment 2, the visual feedback of the writing size was transformed by an enlargement or a reduction relative to the participants’ normal writing trace. In both experiments, on every trial, participants were instructed to produce a particular target size (either large or small) and a particular target slant (either forward or backward) as indicated on the screen, and to stick to the instructed size and slant as close as possible. When an unpredictable transformation occurred, the rescaled visual feedback required adjustments of the subjects, writing movements in order to produce the originally instructed size or slant. It was unknown to the participant whether such a visual feedback transformation was forward or backward, or enlarged or reduced.