Generalized Anxiety Disorder (GAD) is associated with an attentional bias toward emotional information leading to behavioral manifestations which need to be clarified. In healthy participants, the emotional visual stimuli have been shown to capture the attentional resources not only in central vision (CV) but also in impoverished visual conditions as those of the peripheral vision (PV). Hence, it seems necessary to study the processing of emotional information and its behavioral impact in both CV and PV. Objective: The aim of this study was to compare, in patients with GAD and in healthy controls (HC), the behavioral impact of the emotional value of natural scenes presented in CV or in PV. Method: Ten patients with GAD and ten HC matched for age, gender and education level were presented with standardized emotional pictures from the International affective Picture System (IAPS). Three groups of 300 pictures (unpleasant, neutral and pleasant) were presented at three visual eccentricities (CV: 0°; PV: -12° and +12°). The 900 trials were split into 4 blocks during which the participants had to fix a cross in the center of the screen and to report the location of the pictures (left, center or right) using a keypad. We analyzed the behavioral responses: reaction time, the accuracy and the Speed-Accuracy Composite Score (SACS) which measures the efficiency of the response, and the dynamic of the cerebral events using the magnetoencephalography (MEG). Results: The reaction times of patients with GAD were reduced when pictures were emotional while those of HC were increased. Besides, while HC showed a lower efficiency when discriminating the location of pictures appearing in CV, patients demonstrated an equal efficiency in CV and PV. An early occipital component (90 ms) was correlated with the emotional effect. Discussion and conclusion: Patients with GAD showed a high sensitivity to emotional visual information, revealed by an improvement of the efficiency when they responded to emotional pictures and by an early occipital activation. Furthermore, while patients showed similar efficiency regardless of the position of the pictures, HC seemed to expend more attentional resources to monitor the peripheral field which appeared detrimental to the CV. Therefore, our data suggest that a large distribution of attentional resources in visual space is easier for anxious patients than for controls.