Medio-frontal cortex: performance monitoring and beyond

Chair:  Boris Burle,

LNC, Aix-Marseille Univ., CNRS, Marseille, France

Abstract:

In the last fifteen years, there has been an explosion on the investigation
of the role of medio-prefrontal cortex in higher cognitive function, especially
in action monitoring and error processing. Although the participation of the
MPFC in action regulation is now clearly established, several questions remain
un-answered. What are the respective roles of the various areas (Supplemen-
tary Motor Areas, Anterior cingulate cortex etc...) in such regulation? What
are the mechanisms through which behavioral adaptation is performed? What
is the link between MPFC and other structures also involved in executive control
(Lateral part of the frontal cortex, basal ganglia etc...). By coupling different ap-
proaches (surface EEG in time and frequency domains, intra-cerebral recordings
in Monkeys and in Humans), this symposium aims at clarifying some of these
issues. Steinhauser will improve our understanding of the role of the so-called
“Error Negativity” (Ne, or “Error Related Negativity”, ERN), a medio-frontal
negative EEG component peaking immediately after error responses, by show-
ing that it very quickly encodes the valence of the error. The question of the
generator of the Ne will be addressed by Bonini, who will present intra-cerebral
recordings in Humans. Ridderinkhof will provide new insights on how MPFC
might exert its influence on other, lower levels, areas through theta synchroniza-
tion and Procyk, thanks to single cell recording and corticogram in monkeys,
will enlighten the communication between the MPFC and the Lateral pre-frontal
cortex in action regulation.

Talk 1:

Rapid evaluation of error significance during performance monitoring

Marco Steinhauser¹ , Martin E. Maier^1,2,3 & Jonas Matuschek^1
1 University of Konstanz, Konstanz, Germany, ² Università di Bologna, Bologna,Italy, ³ Centro Studi e Ricerche in Neuroscienze Cognitive, Cesena, Italy

The continuous monitoring for errors in ongoing behavior is crucial for achieving goal-directed performance. To adaptively adjust behavior in response to an error, it is not only necessary to detect the occurrence of an error but also to evaluate its significance for future behavior. Although evidence for such an evaluation has been provided, little is known about the speed and flexibility of this process. In the present study, we used event-related potentials to investigate whether error significance influences early correlates of performance monitoring under conditions where significance had to be evaluated during task execution. In two experiments, participants responded to the location of a target stimulus while ignoring two simultaneously presented distractor stimuli that were associated with different amounts of monetary loss. While behavioral responses to low-loss and high-loss distractors were equally frequent, the error-related negativity, a negative deflection peaking immediately after error responses, was larger if the error was associated with a higher loss. This suggests that information about error significance is evaluated during task execution, which implies a rapid and flexible evaluation process. 

Talk 2:

The role of supplementary motor area in action monitoring: evidences from intracerebral ERP recordings in Humans

Francesca Bonini^1,2 , Boris Burle¹ , Catherine Liégeois-Chauvel² , Patrick Chauvel² & Franck Vidal^1
1Laboratoire de Neurbiologie de la Cognition, Aix-Marseille Univ., CNRS, Marseille, France, ² Brain Dynamics Institute, INSERM U751, Marseille, France

Performance evaluation and errors processing are fundamental for adaptive and flexible goal-directed behaviour. Electrophysiological approaches have shown a scalp-recorded event-related potential (ERP), called “error negativity” (Ne) and initially considered to reflect an “error detection” mechanism, while functional neuroimaging, together with some source localization studies, have pointed out the rostral cingulate zone (RCZ) as a possible generator of this activity. Nevertheless the evidence of a similar, even though smaller, ERP on correct trials has challenged the interpretation of Ne’s nature and hence all current models of cognitive control processes. A condition necessary to establish if these two negativities reflect a common functional mechanism is the presence of a common cerebral source. In the present study we first questioned the anatomical source of the Ne, as to date only indirect data are available in humans. Secondly, we tried to assess the unicity of the generator for the negativities on correct and errors, to support the hypothesis of a unique modulated physiological process. Direct recordings from human cerebral cortex in epileptic patients show that supplementary motor areas (SMAs) are implicated in action monitoring. Moreover SMAs are involved in the generation of both error and correct trials evoked responses. These results suggest that these two scalp ERPs do not reflect two distinct brain activities but rather a single process whose amplitude is modulated by performance.    

Talk 3:

The engagement of adaptive control is reflected in oscillatory neural dynamics in mediofrontal cortex

K. Richard Ridderinkhof
University of Amsterdam

The mediofrontal cortex (MFC) is key to adaptive behavior.  Across a variety of situations and paradigms, ranging from reinforcement learning to response conflict and post-error adjustment, the MFC recruits other brain regions to implement and fine-tune such adaptive behavior.  Physiologically, these interactions may occur through local and long-range synchronized oscillation dynamics, particularly in the theta range (4-8 Hz).  Here we report on time/frequency analysis of EEG data from a handful of studies in humans to demonstrate that the MFC-theta signature of such adjustments 1) differs between impulsive errors and attentional lapses, 2) differs between response conflict and stimulus conflict, 3) accurately reflects the temporal dynamics of response conflict, 4) accurately predicts successful learning from negative feedback, and 5) shows qualitative change with age.  These patterns highlight the central role of MFC-theta oscillations in the neurobiological mechanisms underlying the engagement of adaptive control in response to endogenous and exogenous demands.

Talk 4:

Neural correlates of cognitive control modulation during learning in monkeys

Emmanuel Procyk
Stem-Cell and Brain Research Institute, Lyon 1 Univ., INSERM, Lyon, France

Neural unit recordings in the anterior cingulate cortex have shown activity specific of particular outcomes or feedbacks during a trial and error task. This activity, which will be reviewed, represents different events that are all relevant for behavioural adaptation. This includes negative or positive feedbacks after choices, negative feedback after execution errors, or visual signals indicating new conditions. The specificity is suggestive of a mechanism involved in triggering adaptations like shifting after choice errors, compensating after execution errors, etc. In this context we are investigating the correlates of these adaptations in areas directly or indirectly connected to the anterior cingulate cortex.
Neural correlates of adaptation have been studied using unit, LFP, and ECoG recordings in frontal cortex of monkeys performing a trial and error task. In this task monkeys have to find by trial and error in each block of trials which of four targets is rewarded, and then repeat the correct response. The reward schedule is deterministic. The solution is changed after the monkey has repeated at least 3 times the correct response. Animals thus alternate between exploration and exploitation periods that require different levels of control on behaviour.
Our data show changes in lateral prefrontal activity and more distant precentral ECoG oscillations during adaptations. Low frequency (beta) oscillations are modulated by behavioural periods (learning vs repetitive behaviours), and after different feedbacks. These modulations seem to reflect changes in a specific task-related neural process, possibly cognitive control.