Right fronto-parietal involvement in monitoring spatial trajectories

Abstract

This study investigates whether the monitoring role that has been ascribed to the right lateral prefrontal cortex in various cognitive domains also applies to the spatial domain. Specific questions of the study were (i) what kind of spatial contingencies trigger the putative monitoring function of right lateral prefrontal cortex and (ii) which other brain regions are functionally connected to it in monitoring-related conditions. Participants had to track the trajectory of a car moving within a roundabout and detect when the car hit the crash-barrier. Four different trajectories were used with different degrees of regularity and predictability. The results showed that two regions in the right hemisphere, the lateral prefrontal and inferior parietal cortex, were maximally activated and functionally connected when monitoring regular predictable trajectories as compared with unpredictable ones, demonstrating that this fronto-parietal network plays a role in monitoring environmental contingencies that can inform expectancy in a meaningful way.

Introduction

The prefrontal cortex has been traditionally thought as the seat of high-level cognitive operations. Mounting evidence shows functional deconstruction within prefrontal cortex. For instance, recent neuropsychological and neuroimaging studies have shown that left and right lateral prefrontal cortices are relatively more associated with computationally different processes such as criterion-setting and monitoring, respectively (e.g., Alexander et al., 2007, Shallice et al., 2008, Stuss et al., 2002, Vallesi et al., in press; see also Godefroy et al., 1999).

In particular, neuropsychological (Stuss et al., 2005, Triviño et al., 2010, Vallesi et al., 2007a), Transcranial Magnetic Stimulation (TMS; Vallesi et al., 2007b) and neuroimaging (Coull et al., 2000, Vallesi et al., 2009a, Vallesi et al., 2009b) studies have shown that the right lateral prefrontal cortex is important to monitor temporal probabilities. For instance, it plays a role in optimizing behavior when the probability of a target occurrence increases with elapsing time, such as in the variable foreperiod paradigm (Vallesi et al., 2007a, Vallesi et al., 2007b, Vallesi et al., 2009a). In this paradigm, the right dorsolateral prefrontal activation correlates with the RT difference between short and long foreperiods, the latter being associated with a high conditional probability of target occurrence. A monitoring role has been attributed to right lateral prefrontal cortex also in other domains, such as episodic memory retrieval (Henson et al., 1999, Crescentini et al., 2010, Vallesi and Shallice, 2006) and problem solving/reasoning (Reverberi et al., 2005). Although the evidence gathered from different fields and tasks probably advocates a broad monitoring role of right prefrontal cortex, in the context of the present study we use the following operational definition of this process: checking environmental changes that modify the probability of occurrence of critical events, with the goal of optimizing a response to those events.

The aim of the present fMRI study is to test whether not only the right lateral prefrontal cortex but, more extensively, a right fronto-parietal network is involved in monitoring probabilities in a domain different from the temporal one, namely space. We focus on the spatial domain for the following reasons. First, fronto-parietal regions in the right hemisphere are preferentially involved in temporal and spatial predictions (Beudel et al., 2009). Second, visuospatial orientation of attention has been attributed to right superior temporal (Karnath et al., 2004) and inferior parietal regions, such as the supramarginal (Vallar and Perani, 1986) and angular gyri (Mort et al., 2003) in works on unilateral spatial neglect and neuroimaging studies on healthy participants (Corbetta and Shulman, 2002, Galati et al., 2000). A recent TMS study on healthy individuals performing line bisection judgments attributed a more important role to the right supramarginal gyrus than to the right superior temporal or angular gyri (Oliveri and Vallar, 2009). Moreover, lateral prefrontal and parietal regions, which subserve spatially guided behavior, have many common efferent projections in the brain (Selemon and Goldman-Rakic, 1988) and show reciprocal effective connectivity through superior and longitudinal fasciculi. These fiber tracts, in turn, if lesioned in the right hemisphere, may also produce neglect symptoms (Doricchi and Tomaiuolo, 2003, Thiebaut de Schotten et al., 2005).

We therefore expected functional connectivity between right prefrontal and parietal regions, specifically when monitoring of spatial contingencies is advantageous for the behavior. Therefore, a second aim of the study was to assess whether the right fronto-parietal network is specifically involved in monitoring spatial trajectories that are informative about the probability of occurrence of a critical event, rather than in monitoring spatial contexts in general.

To test this hypothesis we designed a visuo-spatial tracking task, in which participants were asked to play the role of “traffic agents” that had to constantly monitor the behavior of an inattentive driver. They had to detect when the driver's car moving within a roundabout hit either the external or the internal crash-barrier (see Methods and Fig. 1 for details). During non-baseline periods, the car moved following one of four different types of spatial trajectories with different degrees of regularity and predictability. In a regular predictable trajectory, for instance, the car progressively approached either the internal or the external barrier until it actually struck the barrier. Predicting the occurrence of an accident by monitoring the spatial trajectory was impossible in the other trajectory types (regular unpredictable, random and zig-zag).

Our main prediction was that a right fronto-parietal network would be more engaged and functionally coupled throughout the highly probabilistic (i.e., regular predictable) trajectories than during the other kinds of trajectories. The latter trajectories were expected to activate the right fronto-parietal network gradually to a less extent, with minimal activation associated to the zig-zag trajectory. In this condition, monitoring processes would in fact be of no help, since approaching a crash-barrier was often misleading because the car then turned back toward the center of the road a variable number of times before hitting one of the barriers.