of upper-alpha power (UAP) neurofeedback training (NFT). A plausible rationale for such training, with an aim to improve cognitive task performance, can be made based upon what is already known of the links between UAP and cognition. However, the potential performance enhancing benefits of UAP-NFT in women and its dependence on hormonal state remain unclear. The aim of this study was to show that neurofeedback training efficiency (TE) is modulated by the menstrual cycle. Methods: The authors collected saliva samples in 3-day intervals over six weeks from 35 young women with a regular menstrual cycle. Menstrual cycle phase was ascertained by saliva progesterone. Women (18–30 years old) were randomly allocated to one of two groups (18-NFT vs. 17-Mock) controlling for age. For excluding novelty of experimental condition influence half of subjects started NFT at menstrual phase (MPS) and half started at mid-luteal phase (LPS) of menstrual cycle. EEG and frontal muscle EMG were recorded before, during and after NFT or Mock session to examine simultaneous upper alpha increase and EMG decrease. The total number of sessions was 10–11. Before and after every NFT session women were asked to perform Kraepelin test feedback of both upper alpha EEG and EMG information which was provided using a simple sound signal and based on thresholds set according to measurements taken during each pre-training resting period. Hence, when power in the upper alpha band increased above that seen at rest with eyes close and power in the EMG signal simultaneously decreased below that seen at rest then a sound was played. TE was calculated as the percentage of time in seconds upper alpha power increased whilst EMG simultaneously decreased during one session. Results: TE of first NFT session was lower in MPs than in LPs (t = 6.9, p = 0.00). Over 10 sessions TE increased in NFT groups [F(9,71) = 5.6, p = 0.011 in MPs and F(9,71) = 4.2, p = 0.023 in LPs] and didn't change in Mock (p b 0.05). Meanwhile TE- increased consequently from MP to LP and decreased from LP to MP in MPs of Mock group. Fluency in cognitive task performance and TE positively correlated with salivary progesterone level. Conclusion: Neurofeedback training efficiency is modulated by the menstrual cycle. The menstrual cycle should therefore be taken into consideration in biofeedback therapy. doi:10.1016/j.ijpsycho.2014.08.670 Neurofeedback games for the enhancement of cognitive abilities related to multitasking Olga Sourina a , Xiyuan Hou a , Lipo Wang b , Wei Lun Lim b , Yisi Liu a a FraunhoferIDM@NTU, Nanyang Technological University, Singapore b School of EEE, Nanyang Technological University, Singapore Various approaches and algorithms are proposed and implemented in neurofeedback systems for training of human cognitive abilities. Neurofeedback training can be done in different forms such as games, simple visual feedback such as color changes, or audio feedback such as a beeping sound. In this work, first, we propose and implement neurofeedback games that can be used with different neurofeedback algorithms to train cognitive abilities related to multitasking such as working memory and attention. Each game consists of a neurofeedback system with implemented neurofeedback algorithms and game flow part. The user has to choose a neurofeedback algorithm (for example, alpha, theta/beta or fractal dimension based training algorithm) through the interface. The user's individual alpha peak frequency and individual alpha band width can be calculated and entered as well to make the training more effective. Second, we propose a novel multitasking game that can be used both as the game for neurofeedback training and as the multitasking test with a score. The score includes the user's performance and reaction time. The player has to drive the car on the road and avoid the walls, react to the sounds in the auditory oddball task, and remember the letters and shoot the target in case the repeated letters appeared. Thus, the multitasking game includes 3 simultaneous tasks involving the player's coordination such as avoiding the walls during the car driving; player's comprehension of the sounds such as recognition of 2 different sounds, and memorization of the letters such that the player has to remember the letters shown on the screen. A neurofeedback training algorithm is integrated in the game as follows: if the targeted EEG parameter is achieved – the driving environment stays without changes; if it is not – the driving environment is shaken, and the color is changed. The multitasking game is more challenging for the player than traditional neurofeedback games because not only the neurofeedback algorithm is used in the game but the simultaneous tasks are integrated as well. Thus, the multitasking game needs to be tested and compared with the traditional neurofeedback games to assess its effectiveness. doi:10.1016/j.ijpsycho.2014.08.671 Individual alpha EEG indices as AD/HD syndrome severity predictors in children Elena Sapina Molecular Biology and Biophysics Research Institute, Russian Medical Science Academy, Russia Background: Attention deficit/hyperactivity disorder (AD/HD) is known as wildly spread neurobehavior disorder. Lately high interest aroused to AD/HD profile personalization (Arns et al., 2008; Barry et al., 2007). Most AD/HD behavior problems are connected with poor self- regulation (Barkley et al., 2010). In previous studies it was shown that self-regulation ability is associated with alpha wave activity. The aim of this study was to examine individual alpha rhythm indices in children with different types and severity of AD/HD. Methods: 47 boys aged 7–11 diagnosed with AD/HD were divided into three groups according to the type of disorder: 19 of combined (C), 15 predominantly hyperactive–impulsive (PHI) and 13 predominantly inattentive (PI) types. AD/HD severity was defined according to test performance: (1) Go/no-Go task (2) modified Stanford test and (3) SNAP-IV. Resting-state EEG was recorded with eyes closed (1 min) and eyes opened (1 min) conditions in Pz. Individual alpha peak frequency, percentage and duration of alpha amplitude suppression in response to eyes open were compared between AD/HD types and groups. ANOVA tests indicated a range of significant difference (p ≤ 0.05) between severity levels of AD/HD. Results: It was shown no difference between AD/HD types in studied individual alpha activity indices. Individual alpha peak frequency was lowest in children with most severe AD/HD. Alpha amplitude suppression was significantly deeper and lasted longer (p b 0.05) in children with medium and light severity (average 1.5–2.4 scores in SNAP IV hyperactive–impulsive and inattentive scales) in comparison with children of severe AD/HD (average 2.5–3.0 scores in SNAP IV hyperactive–impulsive and inattentive scales). Conclusion: The findings suggest that further investigation is warranted to explore the use of individual alpha peak frequency, percentage and duration of alpha amplitude suppression for prognostic indicators of AD/HD severity and its potential to support neurofeedback in rehabilitation programs for children. doi:10.1016/j.ijpsycho.2014.08.672 International Journal of Psychophysiology 94 (2014) 120–261 149