1. Introduction

2. The hypothesis

3. Proposed protocol

4. Implications for practice

References

Abstract

Cyberbullying is a prevalent concern around the world. Research shows that interactions online are associated with similar structural correlates and patterns of brain activity to real-world (offline) relationships, and that the brain experiences peer victimisation (e.g., cyberbullying) in the same way that it experiences physical pain. Furthermore, these experiences can become biologically embedded in the physiology of the developing person, thereby increasing their risk of developing mental health problems. With the increasing prevalence of cyberbullying and youth internet usage, there is a pressing need to further understand the brain’s response to cyberbullying.

We hypothesise that a unique pattern of brain activation is associated with cyberbullying and can be identified using task-based functional magnetic resonance imaging (tbfMRI). However, there is a dearth of research regarding cyberbullying and no fMRI paradigm exists in a real-time situation such as observing a cyberbullying scenario. Here, we propose a tbfMRI protocol we have developed specifically for this purpose

his paper will describe a tbfMRI protocol that can be used to investigate the hypothesis. The overall aim of such a protocol is to elucidate the neurobiological underpinnings of cyberbullying by exploring the brain responses in passive cyber-bystanders (those who witness cyberbullying). This would be the first research to use fMRI to examine brain activation in cyberbystanders, and will bring us closer to understanding the various neurobiological underpinnings that may be associated with cybervictim/bully status and outcomes.

Introduction

Cyberbullying can have serious impacts on mental health (Fahy et al., 2016; Le et al., 2017; McLoughlin, Spears, & Taddeo, 2018; McLoughlin, Spears, Taddeo, & Hermens, 2019), and is commonly defined as an aggressive, repeated, intentional act carried out on an individual using electronic forms (Smith et al., 2008). Cybervictims report significantly more social difficulties, higher levels of anxiety and depression, and are more likely to suffer suicidal ideation (Kowalski, Giumetti, Schroeder, & Lattanner, 2014; van Geel, Vedder, & Tanilon, 2014) than victims of traditional bullying. Prevalence estimates vary, however, The AU Kids Online study (Green, Brady, Ólafsson, Hartley, & Lumby, 2012) found that 29% of Australian children (19% across Europe) said they had been bullied, and 13% of those bullied said this occurred on the internet.

Mild cognitive impairment (MCI) represents the intermediate stage between normal cerebral aging and dementia associated with Alzheimer’s disease (AD). Early diagnosis of MCI and AD through artificial intelligence has captured considerable scholarly interest; researchers hope to develop therapies capable of slowing or halting these processes. We developed a state-of-the-art deep learning algorithm based on an optimized convolutional neural network (CNN) topology called MCADNNet that simultaneously recognizes MCI, AD, and normally aging brains in adults over the age of 75 years, using structural and functional magnetic resonance imaging (fMRI) data. Following highly detailed preprocessing, fourdimensional (4D) fMRI and 3D MRI were decomposed to create 2D images using a lossless transformation, which enables maximum preservation of data details. The samples were shuffled and subject-level training and testing datasets were completely independent. The optimized MCADNNet was trained and extracted invariant and hierarchical features through convolutional layers followed by multi-classification in the last layer using a softmax layer. A decision-making algorithm was also designed to stabilize the outcome of the trained models. To measure the performance of classification, the accuracy rates for various pipelines were calculated before and after applying the decision-making algorithm. Accuracy rates of 99.77% ± 0.36% and 97.5% ± 1.16% were achieved for MRI and fMRI pipelines, respectively, after applying the decisionmaking algorithm. In conclusion, a cutting-edge and optimized topology called MCADNNet was designed and preceded a preprocessing pipeline; this was followed by a decision-making step that yielded the highest performance achieved for simultaneous classification of the three cohorts examined.

Introduction

Cognitive impairment is a general term referring to impairments in cognition among the domains of memory, learning, concentration and decision-making. Cognitive impairment ranges from mild to severe and the symptoms can worsen over time and ultimately prevent a patient from performing daily tasks. Mild Cognitive impairment (MCI) was first utilized by Reisberg et al. [1] and is currently defined as a decline in cognitive ability that is detectable however lacking in terms of the severity to alter one’s functioning of daily living. The National Institute on Aging Alzheimer’s Association (NIA-AA) has provided criteria to diagnose dementia and MCI when there occurs a significant cognitive deterioration from an individual’s previous level [2], [3]. Additionally, research demonstrates that elderly adults with a diagnosis of MCI have a higher risk of developing dementia and age-related cognitive decline [4]. Petersen et al.’s research demonstrates that although there is still a scoring threshold in determining MCI, the memory decline of individuals with MCI is approximately1.5 standard deviations below normative data of same age and educationally matched peers [5], [6]. Gallagher et al. indicate depression and anxiety have been reported in almost 50% of individuals with MCI, and a link between depression and anxiety with cognitive decline has been found [7]–[9].

Attention Deficit Hyperactive Disorder (ADHD) is one of the most common diseases in school aged children. In this paper, we consider using fMRI data with classification techniques to aid the diagnosis of ADHD and propose a bi-objective ADHD classification scheme based on L1-norm support vector machine (SVM). In our classification model, two objectives, namely, the margin of separation and the empirical error are considered at the same time. Then the normal boundary intersection (NBI) method of Das and Dennis is used to solve the bi-objective optimization problem. A representative nondominated set which reflects the entire trade-off information between the two objectives is obtained. Each representative nondominated point in the set corresponds to an efficient classifier. Finally a decision maker can choose a final efficient classifier from the set according to the performance of each classifier. Our scheme avoids the trial and error process for regularization hyper-parameter selection. Experimental results show that our bi-objective optimization classification scheme for ADHD diagnosis performs considerably better than some traditional classification methods.

Introduction

Attention Deficit Hyperactivity Disorder (ADHD) is a very common mental disorder in childhood. ADHD symptoms include inattention, hyperactivity, and impulsively. It affects approximately 5–7% of all school-age children and more than one half of the ADHD children continue to manifest clinically significant symptoms after reaching adulthood [1]. As the pathogenesis is not clear, the main diagnostic method is based on the subjective experience of doctors, which results in many children not being able to receive good treatment in the early stage of ADHD. In addition to the traditional clinical diagnosis, there is a pressing need to find a set of more distinctive and objective features to characterize ADHD that can be used to facilitate ADHD diagnosis. As a promising neuroimaging tool, functional MRI (fMRI) has been widely used to examine the brain of ADHD patients. Abnormal brain activations were found in task-related experiments on the dorsal anterior cingulate cortex (dACC), the ventrolateral prefrontal cortex (VLPFC) and the putamen [2]. Using resting-state fMRI, abnormalities were found in prefrontal cortex, inferior frontal cortex, sensorimotor cortex, anterior cingulated cortex, putamen, temporal cortex and cerebellum. In addition, Castellanos et al. (2008) [3] found ADHD-related decreases of functional connectivity between anterior cingulate and precuneus/posterior cingulate cortex regions, as well as between precuneus and other default-mode network components, including ventromedial prefrontal cortex and portions of posterior cingulate cortex. This suggests that functional connectivity information of functional MRI data can be used as a classification feature for ADHD diagnosis. With the development of machine learning techniques, many efforts have also been made to predict ADHD disease of patients. Mueller et al. [4] have introduced a machine learning system that uses support vector machine (SVM) to differentiate ADHD adults from control groups on the base of the event related potentials that are generated from the EEG measurements. In [5], Peng et al. utilized structural MRI data and ELM for ADHD classification. Dai et al. [6] proposed using multimodal magnetic resonance imaging to classify ADHD children. Mourão-Miranda et al. [7] applied SVM algorithm to perform multivariate classification of brain states from whole fMRI volumes. They demonstrated that SVM outperforms Fisher Linear Discriminant (FLD) classifier in classification performance as well as in robustness of the spatial maps obtained by a comparative analysis. Traditional machine learning classification methods mentioned above for ADHD are based on single optimization technique. One or more hyper-parameters need to be selected before training a classifier. This often leads to time-consuming training sessions and classification inefficiency with changes in sample size. For example, when using a SVM to classify mild cognitive impairment subtypes [8], Haller [9] explored the gamma parameter iteratively from 0.01 to 0.09.

SUMMARY

Purpose: In simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI), increased neuronal activity from epileptiform spikes commonly elicits positive blood oxygenation level–dependent (BOLD) responses. Negative responses are also occasionally seen and have not been explained. Recent studies describe BOLD signal changes before focal EEG spikes. We aimed to systematically study if the undershoot of a preceding positive response might explain the negative BOLD seen in the focus.

Methods: Eighty-two patients with focal epilepsy who underwent EEG-fMRI at 3T were retrospectively studied. Studies with a focal negative BOLD response in the region of the spike field were reanalyzed using models with hemodynamic response functions (HRFs) peaking from )9 to +9 s around the spike.

Results: Eight patients met the inclusion criteria, showing negative BOLD responses in the spike field on standard analysis. None had positive BOLD responses immediately adjacent to the areas of deactivation. Regions of deactivation were found to have congruent preceding positive responses in two cases. These early activations were seen at the combined maps of )5 to )9 s.

Discussion: This study indicates that in a small proportion of patients with focal epilepsy in whom the standard analysis reveals focal negative responses, an earlier positive BOLD response is probably the cause. The origin of negative BOLD signal changes in the focus as a result of an epileptic event remains, however, unexplained in most of the patients in whom it occurs. KEY WORDS: fMRI, Focal epilepsy, Spike, Deactivation.

Currently, there is still not enough understanding of how the human brain perceives the information input from the outside world and how neurons react correspondingly in a conscious and/or unconscious manner of perception. Non-invasive brain activity recording technologies, like electroencephalography (EEG) and functional magnetic resonance imaging (fMRI), have been widely introduced to record human brain dynamics under certain circumstances. EEG measures brain activities over time with a high temporal resolution of milliseconds, while fMRI mainly identifies which area of the brain is in use with a high spatial resolution of millimeters. There is a convergent evidence that suggests that brain activity recordings play a vital role to boost the development of new biometric technologies and precede studies of brain functions like attention and memory (Han et al., 2015), motor control (Heimann, Umiltà, Guerra, & Gallese, 2014), and emotions (Alarcao & Fonseca, 2017). In the fields of brain encoding and decoding, there has been a number of studies over the past decades addressing one of the basic questions of how information is represented in the brain (Naselaris, Kay, Nishimoto, & Gallant, 2011). As a consequence, the connections between brain activities in the visual cortex and lowlevel visual features such as orientation (Haynes & Rees, 2005), color (Brouwer & Heeger, 2009), and position (Thirion et al., 2006) have been intensively studied. To measure the brain responses while watching natural images, Kay, Naselaris, Prenger, and Gallant (2008) proposed an fMRI-based decoding system to apply a receptive field-based model to represent individual fMRI voxels. They modeled a generation process of fMRI signals in particular in visual areas V1, V2 and V3, and further tested the performance in an application of image identification. A high identification accuracy was obtained from two participants, suggesting the feasibility to predict novel natural images by using the proposed general visual decoder. Following this previous study, Naselaris, Prenger, Kay, Oliver, and Gallant (2009) proposed a Bayesian framework to model fMRI signals, and were successful in reconstructing natural images based on fMRI. According to their method, individual voxels were modeled in two different approaches, namely the Gabor wavelet-based structural encoding model and semantic-based encoding model, thus characterizing the fMRI responses in the early visual areas and anterior visual areas, respectively. Instead of using static images as visual stimuli, Nishimoto et al. (2011) presented a motion energy-based encoding model to represent the fMRI signal patterns in the early visual areas while watching natural movies, and demonstrated the validity and temporal specificity of this encoding model. Furthermore, they applied a Bayesian decoder to test the reconstruction accuracy of using brain activity measurements to reconstruct the dynamic visual information in movies. Similarly, Han, Zhao, Hu, Guo, and Liu (2014) introduced an fMRI-based encoding model to predict the brain network response while the participants were free-viewing video clips. They pointed out that a successful brain encoding technique could benefit the evaluation and guidance of visual feature extraction in applications of visual attention and image processing, and could also boost the development of cognitive neuroscience studies.

Consumer commodity sector spends huge amount of money on advertising the usability or success of the products. This is necessary during pretesting of various alternative advertisement campaigns before launching and also during the in-market analysis of the campaign after launching. A number of conventional methods exist to pretest advertisements. These methods include self-reported methods such as liking, recall or purchase intent, neuromarketing, etc. “Neuromarketing” phrase is the combination of two words “neuro” and “marketing”. This can be referred to as merging of two fields, namely “neuroscience” and “marketing”. Using neuroscience, it is possible to improve the marketing of the existing products. It provides an insight to improve the design of products before actual launch in the market [24]. Therefore, neuroscience reveals information about consumer preferences on products. This is not possible using traditional methods such as questionnaire, attitude or verbal communication. Thus, neuromarketing can help the product-companies to reduce expenditure on product advertisements. Over the past few years, researchers have developed different neurophysiological methods for analyzing the consumer behavior and advertisement phenomenon to study different aspects of marketing [7, 20]. The growth has been recorded due to the technological advancement of instruments such as Functional magnetic resonance imaging (fMRI), Magnetoencephalography (MEG), and electroencephalography (EEG) often used in neuromarketing. fMRI is a neuroimaging technique used to measure the amount of deoxygenated hemoglobin during neuronal activity. It describes brain function with spatial and temporal resolutions of one millimeter and per second change, respectively [33]. The device has been used for grading user preferences on various applications. Berns et al. [6] have used fMRI data to predict music popularity by measuring the brain activity of 27 adolescents. The authors have shown that the positive correlation of the participant’s brain responses with the sales after three years. However, the cost associated with such experiments can be very expensive and often turns out to be in the order of several millions of dollars [35]. Moreover, additional costs such as insurance, maintenance and portability issues limit the use of fMRI in such applications. MEG is used to capture the brain activity that is caused by neuron activity. Such neuron activity creates a magnetic field that is amplified and mapped by MEG with spatio-temporal information. However, due to higher cost, technical complexities and limited access of subcortical regions for imaging, it is not preferred as a device to develop systems of neuromarketing.

The growth of social problems resulting from the deterioration of the environment calls for changes in the behavior of members of society. Actions that are environmentally desirable such as lowering greenhouse gas emissions, reducing waste, and increasing clean energy and water, can only be met through higher levels of public participation (Brunson and Reiter, 1996; Mckenzie-Mohr, 2000). An approach with an outstanding potential to foster sustainable behaviors is the design of information-intensive advertising campaigns (VanDyke and Tedesco, 2016). Information offered to members of society or potential ecological consumers through advertising enhances awareness of environmental issues (FrajAndres and Martínez-Salinas, 2007
), encourages development of attitudes and intentions bolstering green consumption (Connell et al., 2014) and can even alter behavior (Han et al., 2010). Therefore, if designed properly, environmental messages can shape attitudes and decisions, and lead them in the direction of more responsible behaviors. Given the importance to a proper design of messages, most ecological advertising literature has focused on media features such us gain/loss framing (Tu et al., 2013) or message ambiguity (Leonidou et al., 2011) on attitudes, intentions or behaviors. A media feature typical of environmental messages is temporal framing, a concept that in this case refers to the display of an ecological message using a specific reference to time (Chandran and Geeta, 2004). In environmental communication, a future frame (FF) reveals the consequences of acting in the future for/against the environment, while a past frame (PF) emphasizes the consequences of having acted for/against the environment (Antes and Mumford, 2009). Despite the importance afforded by many studies to future framing (Kees, 2011; Xu et al., 2015), environmental communication research has advanced a variety of findings questioning which type of framing is more persuasive (Martin et al., 2011). Ecological messages are habitually pronounced by voices differing in gender and age. While voice gender (male/female) has gained much interest in communication research, less attention has been paid to the question of whether the age of the voice, young (YV) or old (OV), plays a persuasive role and the current findings regarding this aspect are inconsistent (Mammarella et al., 2013; Zaske et al., 2013 € ).