Gerard P. van Galen

Handwriting: Issues for a psychomotor theory

Abstract In this article it is argued that handwriting basically is a multi-component task. This implies that the production of writing strokes is the overt manifestation of divergent cognitive, psychomotor and biophysical processes. Based upon a review of elementary psychomotor findings and theoretical issues related to the cognitive structure of the skill, a possible architecture for handwriting processes is sketched. In this handwriting model each process has a characteristic unit of processing; receives its input from the operation next higher in the hierarchy; and is responsible for a specific transformation of that information to make it appropriate as an input to the next lower process. To accommodate for processing time frictions between modules, each of them is assumed to have a provision for a transient storage of output. The parallel feature of the model involves that all processors operate concurrently, but on different features of the message.

FITTS LAW AS THE OUTCOME OF A DYNAMIC NOISE FILTERING MODEL OF MOTOR CONTROL

Abstract Studies into the real-time evolution of goal-directed movements have been strongly dominated by the view that movement time and other kinematic features are a direct reflection of on-line computational processes. Well-known examples of this position try to explain the logarithmic relation between the movement time of aiming movements and their demanded endpoint accuracy, as described by Fitts in 1954, by the use of centrally controlled servomechanisms and submovements. In the present article it is doubted whether such a strict cognitive approach can provide real understanding of the kinematics of aiming movements. An alternative, noise filtering model of motor control is presented proposing that the psychomotor system is an inherently noisy mechanical system for which spatial demands should be formulated in terms of a desired signal-to-noise ratio between goal-related propulsion of the limb (signal) and stochastic error (noise). Adequate movements would, from this point of view, result from the optimization between the application of muscle forces to the limb system and the noise reducing effects of biomechanical properties, such as stiffness, viscosity, or friction due to surface contact. At a theoretical level, the present approach is exemplified in a simulation model which takes into account the stochastic nature of the motor unit recruitment process and the noise filtering properties of a biomechanical limb. Empirical data acquired in a simulation study, as well as published data on eye-ball control during looking tasks and data on axial pen pressure control in graphic tasks, lend support to the view that adaptive control of muscular co-contraction is a relevant degree of freedom for the control of spatial accuracy.

Error, stress and the role of neuromotor noise in space oriented behaviour.

In this article both movement errors and successful movements are considered to be the product of varying ratios of muscle force signals and the composite of neuromotor noise in which the force signal is embedded. Based on earlier work we derived four propositions, which together form a theoretical framework for understanding the incidence of error in conditions of time pressure and mental load. These propositions are: (1) motor behaviour is an inherently stochastic and therefore noisy process; (2) biophysical, biomechanical and psychological factors all contribute to the level of neuromotor noise in a movement signal; (3) endpoint variability of movement is related to the signal-to-noise ratio of the forces which drive the moving limb to the target; and (4) optimal signal-to-noise ratios in motor output can be arrived at by adjusting limb stiffness. In an experiment with a graphical aiming task in which subjects made pen movements to targets varying in width and distance, we tested the prediction that time pressure and dual task load would influence error rates and movement noisiness, together resulting in biomechanical adaptations of pen pressure. The latter is seen as a manifestation of a biomechanical filtering strategy to cope with increased neuromotor noise levels. The results confirmed that especially under time pressure error rates and movement noise were enhanced, while pen pressure was higher in both conditions of stress.

Programming in handwriting: Reaction time and movement time as a function of sequence length

Abstract In order to test whether the model developed by Sternberg et al. (1978b) on the programming of speech and typewriting movements also holds for handwriting, three experiments were conducted in which four trained subjects each wrote a few letters in a simple reaction paradigm. In experiment 1, one to four letters had to be written in each of three conditions: normal, large and with the addition of extra strokes. In experiment 2, digits were used, written either in the normal way or according to a novel movement pattern. In experiment 3, the letters formed words or nonsense strings. In all three experiments a linear increase of reaction time with sequence length was found, the slope of which rapidly decreased with practice, however. The time to write a letter did not increase with sequence length. The variables of experiments 2 and 3 had large effects on movement times but no effect on reaction times. A parallel processing model is suggested in which only the first unit is completely preprogrammed before the go signal and in which succeeding elements are programmed during the execution of preceding ones.

The Acquisition of Skilled Handwriting: Discontinuous Trends in Kinematic Variables

The inference of movement control strategies from recorded writing signals is illustrated by presenting the recordings and corresponding absolute velocity profiles of the first attempts in writing an unfamiliar grapheme by an adult subject. A number of kinematic variables derived from these recordings are discussed and used as dependent variables in an experiment. Primary school children from grade two to six wrote the letters of a cursive alphabet in a paced writing task. Movement time, absolute velocity, curvature and writing dysfluency measurements of produced grapheme segments were measured. The changes in these kinematic variables during primary school revealed a discontinuity in writing development. The results show that during the acquisition process a temporary decline in performance measures can be observed as a result of feedback-controlled movement strategies. During the learning of the handwriting skill a feedback-controlled movement strategy is gradually replaced by an open-loop movement strategy.

The independent monitoring of form and scale factors in handwriting

Abstract In an experimental study on movement latencies and durations of letter-writing tasks we tested a three-stage model of the retrieval and initiation of handwriting movements. In this model it is assumed that the process of preparation of handwriting movements consists of the retrieval of a non-muscle-specific motor program in stage one, the substitution of actual parameters into that program in stage two, and finally, the recruitment of an appropriate number of motor units for the execution of the program in the anatomical and physical context in stage three. The pattern of additivities and interactions in the latency data, due to variations of letter size, slope of the base line, and sequencing of letter strokes, was consistent with our three-stage model of complex motor behavior. Furthermore, the analysis of movement times for separate letter strokes revealed a process that involves the unpacking of later strokes during the execution of earlier ones.

Invariants in Handwriting: The Information Contained in a Motor Program

Publisher Summary This chapter presents an experiment that demonstrates a method for determining whether the spatial characteristic or the temporal characteristic of a handwriting pattern is more invariant over replications under normal conditions and under voluntary transformations or extraordinary execution conditions. This method employs three criterions. The first criterion employs the signal-to-noise amplitude ratio of a characteristic. The signal is the average characteristic, while its noise comprises the fluctuations between the average and a specific replication. The second criterion employs intercharacteristic correlation coefficients. The third criterion is concerned with the robustness of a characteristic (expressed by the intercondition correlation) across various instructed global transformations of the movement pattern, such as writing at a different size or speed, or with respect to some arbitrary manipulation of the writing conditions, such as writing without visual feedback or writing on a low-friction surface while reducing proprioceptive feedback. The experimental results consistently showed that the spatial structure was more invariant than the temporal structure. The apparent findings of invariant temporal characteristics in handwriting movements may be because of the motor system that produces the smoothest trace to satisfy the desired spatial outcome. Temporal invariance may also be derived from spatial invariance.

Preparation of partly precued handwriting movements: The size of movement units in handwriting

Abstract In the present study we investigated whether individual strokes or complete letters form “movement units” in cursive handwriting. In various reaction time (RT) paradigms (choice-RT, precue-RT, and simple-RT) we tested which definition of a unit provides the best explanation for the reaction-time and movement-duration data that we observed. In the choice-RT condition we found that congruence of complete letters facilitated reaction time, but congruece of strokes within letters did not. This was also found to hold in the precue conditions where, some time prior to the imperative stimulus, a precue was presented, specifying either the first or the second letter of the writing pattern. Furthermore, analysis of movement durations revealed that the strokes immediately preceding and following the connection stroke between two identical letters were delayed. These results consistently point towards the notion that the movement pattern of a well-practised letter is handled as a single unit.

On the Simultaneous Processing of Words, Letters and Strokes in Handwriting: Evidence For a Mixed Linear and Parallel Model

Publisher Summary This chapter presents a study that traced the time boundaries of processing of motoric demands at different levels of the handwriting process. The chapter illustrates a model of mixed linear and parallel processing motoric demands at two levels: sequential processing at the micro level and parallel processing at the macro level. The hierarchical nature of the model attempted to predict which specific elements of the writing tasks were handled earlier by higher rather than by lower levels of information processing. Specifically, effects of word complexity manifested themselves before those of letter complexity, which, in turn, preceded those of letter connections. Velocity variables and time variables are also discussed in the chapter. Results supported both the existence of level-specific time boundaries and the hierarchical nature of the levels. Writing, by virtue of the slower real time execution speeds of handwriting, allows a more detailed prediction of the real-time execution of hierarchical levels and provides a better opportunity to study motor behavior than speech production.

Perceptual-Motor Complexity of Printed and Cursive Letters

AbstractA number of factors determining the perceptual-motor complexity of letter forms are discussed. An experiment is reported in which primary school children wrote the lower-case letters of a cursive alphabet twice, once after the visual presentation of printed letters and once after the visual presentation of cursive stimulus letters. Response-initiation-time differences between these two types of experimental trials were considered to reflect a cognitive translation process from the graphemic to the motoric level. The analyses revealed that spatial ambiguity, allographic variability, contextual ambiguity, and letter frequency are determinants of the time needed by children for perceiving printed and producing corresponding cursive letters. The motoric complexity of writing cursive letters was investigated by analyzing writing velocity, dysfluency, and curvature measurements of produced grapheme segments. The analyses indicated that letter frequency and the curvature of grapheme segments determine th...