The aim of this thesis is to investigate the isolated and jointed influence of contextual factors,
such as expectancy, conflict and motivation, on the latency distribution of saccades. The underlying
hypothesis is that the analysis of the orienting eye movements recorded from human
volunteers during simple saccadic tasks might provide a good insight into the decision-making
mechanisms of our brain.
The main focus is on a manipulation, to which we refer as the probability-bias manipulation, designed to study the effects, on the generation of saccades, of spatiotemporal uncertainty
regarding the location of the saccade target. Uncertainty is likely to affect the expectancy for a
given sensorimotor event and this is reasonably a crucial determinant for the decision-making
process leading to a saccade. The first central question addressed in chapter 3 is whether and how
probabilistic information about the location of the target is used (consciously or unconsciously)
by the oculomotor system under different task demands. The probability-bias manipulation can
also be seen as a particular instantiation of sequence-related effects, or, in other terms, of those
effects resulting from the history of the past trials. Therefore, the effect of the main manipulation
might interact with the local trial-sequence structure, i.e. the specific trials preceding the current
one. In chapter 4 we present some results about this type of effects.
The use of probabilistic information, and the corresponding adaptation of behaviour to the
stimulus statistics, can be compared to the theoretical limits due to the finite size of the available
sample of data. These limits can be defined in a mathematically precise form on the basis of
the theory of statistical inference and of the Bayesian rule. In chapter 5 we address the issue of
how efficiently, with respect to the theoretical limits, the distribution of saccadic latency adapts
to the probability-bias manipulation. The idea about this type of analysis came from several
interesting discussions with professor William Bialek, at the University of Princeton, during a
short-term visit which was part of my PhD programme.
The interaction of the probability-bias manipulation with the graded involvement of voluntary
control in the generation of saccades is also a main focus of this thesis. For this reason
we tested our subjects with a set of paradigms demanding a different degree of voluntary commitment.
In chapter 6, we analyse in particular the interaction of the probability bias with
a decisional conflict, between the tendency to orient gaze to a visual onset and the required
inhibition of the saccade.
Finally, if we assume that the mechanisms of sensorimotor integration are capable to efficiently
use the probabilistic information about the stimuli to optimise the behavioural performance,
an interesting possibility is that such capability would maximally emerge when the
target stimuli are associated to a strong motivational valence. This issue is examined at several
points in the text of this thesis, and is discussed with particular attention in chapters 3 and 5.
In chapter 7 we analyse the effect exerted by a manipulation of the valence per se, dissociating
it from the probability bias manipulation. In other terms, we analyse the influence that the
motivation to orient to the saccadic goal has on simple stimulus-elicited saccades.
