Psychological Research (2003) 67: 10–21 DOI 10.1007/s00426-002-0094-5 ORIGINAL ARTICLE Ju¨ri Allik Æ Mai Toom Æ Aavo Luuk Planning of saccadic eye movements Received: 9 July 2001 / Accepted: 18 March 2002 / Published online: 21 September 2002 Ó Springer-Verlag 2002 Abstract Most theories of the programming of saccadic completely specified before the movement begins. Di- eye movements (SEM) agree that direction and ampli- rection and amplitude make up a minimal set of inde- tude are the two basic dimensions that are under control pendent dimensions whose values determine the when an intended movement is planned. But they dis- identities of all possible movements. One question that agree over whether these two basic parameters are remains disputed is whether these two basic parameters specified separately or in conjunction. We measured of SEM, direction and amplitude, are specified sepa- saccadic reaction time (SRT) in a situation where in- rately or jointly in a unitary fashion. Neurophysiological formation about amplitude and direction of the required explanations are usually inclined towards holistic models movement became available at different moments in without any distinct computation of direction and am- time. The delivery of information about either direction plitude values (Sparks, 1988). Neurons in the frontal eye or amplitude prior to another reduced duration of SRT fields and in the intermediate and deep layers of the demonstrated that direction and amplitude were speci- superior colliculus, which activates immediately before fied separately rather than in conjunction or in a fixed the onset of a SEM (Robinson, 1972; Sparks, 1978; Lee, serial order. All changes in SRT were quantitatively Rohrer, & Sparks, 1988; Sparks, 1988; Schall & Hanes, explained by a simple growth-process (accumulator) 1993), are arranged topographically to give rise to a model according to which a movement starts when two motor map in which all movements are coded in terms of separate neural activities, embodying the direction and the desired final position of the eye or a fixed movement amplitude programming, have both reached a constant vector, not in separate terms of the direction and am- threshold level of activity. Although, in isolation, the plitude to be moved (Robinson, 1972; Glimcher & amplitude programming was faster than the direction Sparks, 1992). A saccade is produced when the neurons programming, the situation reversed when two dimen- at one location within the motor map become suffi- sions had to be specified at the same time. We conclude ciently active (Ottes, et al., 1984; Schall, 1995; Findlay & that beside the motor maps representing the desired final Walker, 1999). position of the eye or a fixed movement vector, another However, it is questionable that completely specified processing stage is required in which the basic parame- motor programs exist for all possible SEM included in ters of SEM, direction and amplitude, are clearly sepa- the motor map. It is also plausible that the motor map rable. representing the desired final position of the eye is only one of the stages in the programming of an eye move- ment followed (or preceded) by some other stage(s) in which two basic parameters of SEM, direction and Introduction amplitude, are clearly separated from one another. In- deed, psychological data demonstrate that these two Saccadic eye movements (SEM) last only tens of milli- basic parameters of SEM are to a certain extent sepa- seconds, which means that their parameters must be rable from one another. For example, the time required to reprogram a saccade in response to a pair of target displacements, followed one after another, depends on J. Allik (&) Æ M. Toom Æ A. Luuk the spatial relations between these two displacements Department of Psychology, University of Tartu, (Wheeless et al., 1967). If the second displacement is in Tiigi 78, Tartu 50410, Estonia E-mail: [email protected]; the opposite direction to the first target displacement, Tel.: +372-7-375905; the latency of the response is about 40–50 ms longer Fax: +372-7-375900 than the average reaction time to a single displacement 11 alone (Wheeless et al., 1967; Komoda et al., 1973; Hal- seminal paper, Rosenbaum (1980) recommended, as an lett & Lightstone, 1976; Hou & Fender, 1979). It was obvious extension of the precuing technique, to vary the also shown that this extra time is required only if the delay between the precue and movement instruction. new saccade is not in a direction similar to the canceled Unfortunately, as far as we know, nobody has pro- one. If directions of the new and previous saccade co- ceeded in this promising direction. Another limitation of incide and only its magnitude must be corrected, the the precuing technique is the use of a separate cue in reaction time could be even shorter than the average addition to the movement instruction itself. This means, response time to a single displacement. These results in particular, that it is necessary to develop a separate seem to suggest that if one of the two parameters of the explanation to the cue processing besides that of the subsequent displacement remains unchanged there is no model of motor programming itself. For example, it is need to ‘‘rewrite’’ this part of the motor program. As a well documented that an additional signal preceding the consequence, the response time to the second instruction movement instruction can substantially reduce the re- will be shorter in proportion to the amount of time that sponse time even if it contains no specific information is required to specify this particular parameter of the about parameters of the following movement (Saslow, motor program. This means that programming direction 1967). In order to overcome or at least relax some of and amplitude involve separate processes that can be these limitations we developed a new method – the manipulated independently from one another. Although movement parameters disassociation technique – which this interpretation seems rather plausible it is still based can be considered as an extension of the precuing on a rather intricate chain of inferences about how two method. subsequent responses, the canceled and the new one, The basic idea of this new method is to separate the interact with each other. It is also not the only possible programming of the direction and amplitude by pre- way to explain the double-displacement data (cf. Ottes, senting the information about their values separately at van Gisbergen & Eggermont, 1984; Clark, 1999). different time moments. For this purpose, the instruc- Therefore, it is necessary to have a more straightforward tion about required SEM was not presented by indi- method which allows the observation of preparatory cating the exact location, as is customary in this type of processes that precede a single SEM. experiment, but by presenting two symbols; the first One promising candidate is the movement precuing indicating the direction in which to move and the second technique developed by Rosenbaum (1980). The basic the distance through which the eye must move. Dividing idea of the precuing technique is to present an additional the instruction between two separate signals allows them cue before the movement instruction, which gives in- to be presented separately at different time moments formation about some of the spatial parameters of the with a certain Instruction Onset Asynchrony (IOA) be- movement that must be executed on that trial. Regard- tween them. Presenting information about direction less of the informativeness of the precue, the subject before amplitude, or vice versa, provides advance partial cannot produce the required response until the instruc- information and consequently extra time for preparation tion is finally presented eliminating uncertainty about all of the movement attribute that was presented earlier. parameters of movement. The major assumption of this Another way to describe this method is in terms of method is that the subject uses the precued information response alternatives. Before presentation of either of to partially program upcoming movement before the the two signals the total number of alternatives is equal instruction is presented. Comparing precued conditions to the product of the number of potential directions and with uncued ones (the former are typically shorter than the number of potential amplitudes. After the presen- the latter) it is possible to infer how much time it will tation of the first signal the number of alternatives will take, on average, to specify values in the motor program be reduced by the factor equal to the number of alter- that have been precued. Although the rationale of the natives values the first signal has. Thus, the experimental precuing technique is rather simple and transparent, it idea is to vary the time interval with the reduced number has seldom been applied to the study of SEM planning. of response alternatives before the complete elimination In one of the few studies in which the precuing technique of spatial uncertainty. was used, it was shown that subjects are faster in initi- The most important advantage of this new method, ating saccades when they know either the direction or over that of the static movement precuing technique, is amplitude of the required movement in advance, com- that it allows the time course of programming to be pared with a condition without prior knowledge of the observed. In this particular respect, this method is sim- movement parameters required to execute (Abrams & ilar to the timed response paradigm (Schouten & Becker, Jonides, 1988). These results were interpreted as an in- 1967; Ghez, Hening & Favilla, 1990) which was invented dication that the direction and amplitude are specified for the disassociation of the mechanisms triggering separately, and not in a fixed serial order or in con- movement initiation from those specifying response junction. features. By instructing subjects to initiate responses in Beside obvious advantages the movement precuing synchrony with temporally predictable signals and pre- technique also has some limitations.