Revisiting Fitts and Peterson (1964): Width and amplitude manipulations to the reaching environment elicit dissociable movement times.

Abstract

The classic theorem of Fitts (1954) asserts that the combined effects of movement amplitude and target width (index of difficulty: ID) define movement times (MTs) for goal-directed reaches. Moreover, Fitts' theorem states that reaches yielding the same ID produce equivalent MTs regardless of the response's amplitude and width combination. However, most work providing direct support for Fitts' theorem has employed short movement amplitudes and small target widths. Thus, no direct evidence supports the unitary nature of MT/ID relations across a range of amplitudes and widths used in contemporary studies of goal-directed reaching. To that end, we contrasted MT/ID relations for discrete reaches equated for movement ID but differing with respect to their amplitude (15.5, 19, 25.5, and 38 cm) and width (2, 3, 4, and 5 cm) requirements. Results show that amplitude and width manipulations yielded robust linear MT/ID relations; however, the slope of the MT/ID function was markedly steeper in the former (amplitude=92 ms; width=13 ms). Such findings indicate that the constituent elements of movement ID are dissociable and that the fixed parameter nature of Fitts' theorem cannot be applied to a continuous range of veridical movement amplitudes and target widths.

Augmented feedback influences upper limb reaching movement times but does not explain violations of Fitts' Law

“…In what is a salient example for the present study, Heath et al (2011) found that changes in ID induced by an amplitude-based manipulation resulted in a greater influence on MT than changes induced by a width-based manipulation. Specifically, participants performed discrete upper-limb reaches to virtual targets under a one-way mirror wherein concurrent visual feedback of the moving limb was provided via a light emitting diode (i.e., and LED) attached to the forefinger.…”

“…However, contemporary research has reported that amplitude-based contributions to IDs yield larger increases in MTs than width-based contributions. This discrepancy may relate to the presence of augmented terminal feedback in Fitts' original research, which has not been provided in more recent investigations (e.g., Heath et al, 2011). To address this issue, participants performed reaching movements during two sessions wherein feedback regarding terminal accuracy was either provided or withheld.…”

“…Although Fitts' theorem has become prominent in a variety of fundamental and applied research fields, it has also faced criticism (e.g., Welford et al, 1969; Hoffmann and Sheikh, 1991; Danion et al, 1999; Adam et al, 2006; Heath et al, 2011). In what is a salient example for the present study, Heath et al (2011) found that changes in ID induced by an amplitude-based manipulation resulted in a greater influence on MT than changes induced by a width-based manipulation.…”

“…The forefinger of participants was also tracked via a 3D optoelectric system while performing reaches to targets of varying widths (20, 30, 40, and 50 mm) and amplitudes (155, 190, 255, and 380 mm). Notably, Heath et al (2011) employed a “width manipulation” wherein the different target widths were used across a common amplitude (i.e., 255 mm), and an “amplitude manipulation” wherein the different amplitudes were combined with a single target width (i.e., 30 mm). The resultant IDs across width and amplitude manipulations were 3.36, 3.67, 4.08, and 4.67 bits of information.…”

Bimanual and Unimanual Convergent Goal-Directed Movement Times

“…1H D onehanded movements; P D probe held in preferred hand; NP D probe held in nonpreferred hand. distance to a target when all ID values are changed only with a change of target width, such as in some of the experiments of Heath, Weiler, Marriot, Elliott, and Binstead (2011) and Munro, Plumb, Wilson, Williams, and MonWilliams (2007). As shown by Hoffmann and Chan (2012), the geometrical forms of the square root and logarithm terms are nearly identical (r D .99 in these experiments) and hence the regression in terms of log(A) and log(W) can be rewritten in the form of Equation 2.…”