This doctoral dissertation deals with causal modelling in the social sciences. The specific question addressed here is: what is the notion, or the rationale, of causality involved in causal models? The answer to that epistemological query emerges from a careful analysis of the social science methodology, of a number of paradigmatic case studies and of the philosophical literature. The main result is the development of the rationale of causality as the measure of variation. This rationale conveys the idea that to test – i.e. to confirm or disconfirm – causal hypotheses, social scientists test specific variations among variables of interest. The notion of variation is shown to be embedded in the scheme of reasoning of probabilistic theories of causality, in the logic of structural equation models and covariance structure models, and is also shown to be latent in many philosophical accounts. Further, the rationale of causality as measure of variation leaves room for a number of epistemological consequences about the warranty of the causal interpretation of structural models, the levels of causation, and the interpretation of probability. Firstly, it is argued that what guarantees the causal interpretation is the sophisticated apparatus of causal models, which is made of statistical, extra-statistical and causal assumptions, of a background context, of a conceptual hypothesis and of a hypothetico-deductive methodology. Next, a novel defence of twofold causality is provided and a principle to connect population-level causal claims and individual-level causal claims is offered. Last, a Bayesian interpretation of probability is defended, in particular, it is argued that empirically-based Bayesianism is the interpretation that best fit the epistemology of causality here presented. The rationale of variation is finally shown to be involved or at least consistent with a number of alternative accounts of causality; notably, with the mechanist and counterfactual approach, with agency-manipulability theories and epistemic causality, with singularist accounts and with causal analysis by contingency tables.

Schizophrenia is a complex and severe psychiatric disorder with positive symptoms, negative symptoms and cognitive deficits. Preclinical neurobiological studies showed that alterations of dopaminergic and glutamatergic neurotransmitter circuits involving the prefrontal cortex resulted in cognitive impairment such as working memory. Functional activation and functional connectivity findings of functional Magnetic Resonance Imaging (fMRI) data provided support for prefrontal dysfunction during fMRI working memory tasks in individuals with schizophrenia. However, these findings do not offer a neurobiological interpretation of the fMRI data. Biophysical modelling of functional large-scale networks has been designed for the analysis of fMRI data, which can be interpreted in a mechanistic way. This approach may enable the interpretation of fMRI data in terms of altered synaptic plasticity processes found in schizophrenia. One such process is gating mechanism, which has been shown to be altered for the thalamo-cortical and meso-cortical connection in schizophrenia. The primary aim of the thesis was to investigate altered synaptic plasticity and gating mechanisms with Dynamic Causal Modelling (DCM) within functional large-scale networks during two fMRI tasks in individuals with schizophrenia. Applying nonlinear DCM to the verbal fluency fMRI task of the Edinburgh High Risk Study, we showed that the connection strengths with nonlinear modulation for the thalamo-cortical connection was reduced in subjects at high familial risk of schizophrenia when compared to healthy controls. These results suggest that nonlinear DCM enables the investigation of altered synaptic plasticity and gating mechanism from fMRI data. For the Scottish Family Mental Health Study, we reported two different optimal linear models for individuals with established schizophrenia (EST) and healthy controls during working memory function. We suggested that this result may indicate that EST and healthy controls used different functional large-scale networks. The results of nonlinear DCM analyses may suggest that gating mechanism was intact in EST and healthy controls. In conclusion, the results presented in this thesis give evidence for the role of synaptic plasticity processes as assessed in functional large-scale networks during cognitive tasks in individuals with schizophrenia.

Modelling neuronal interactions using a directed network can be used to provide insight into the activity of the brain during experimental tasks. Magnetoencephalography (MEG) allows for the observation of the fast neuronal dynamics necessary to characterize the activity of sources and their interactions. A network representation of these sources and their connections can be formed by mapping them to nodes and their connection strengths to edge weights. Dynamic Causal Modelling (DCM) presents a Bayesian framework to estimate the parameters of these networks, as well as the ability to test hypotheses on the structure of the network itself using Bayesian model comparison. DCM uses a neurologically-informed representation of the active neural sources, which leads to an underdetermined system and increased complexity in estimating the network parameters. This work shows that inform- ing the MEG DCM source location with prior distributions defined using a MEG source localization algorithm improves model selection accuracy. DCM inversion of a group of candidate models shows an enhanced ability to identify a ground-truth network structure when source-localized prior means are used. / Master of Science

Screen media activity such as watching TV and videos, playing video games, and using social media has become a popular leisure activity for children and adolescents. The effect of screen time has been a highly debated topic; however, there is still very little known about it. Using a dataset from the Adolescent Brain Cognitive Development longitudinal study 4 217 young adolescents, that met the requirements, could be retrieved for this thesis project after processing of the data. This thesis project investigated causal order between genetic effect (cognitive performance Polygenic scores (PGSs)), screen time activity, brain morphology (structural Magnetic Resonance Imaging (sMRI) for surface area and cortical thickness), lack of perseverance, and cognitive performance (crystallized IQ) with a machine learning algorithm DirectLiNGAM. A clear correlation between screen media activity and PGS was found for all types of screen time activities but only video games and social media correlated to the global surface area. Furthermore,  TV and video seem to affect lack of perseverance, and lack of perseverance, in turn, affects time spent on video games. These findings imply that different types of social media are not as alike as we thought and can affect adolescents differently. Taken together, these findings support previous research on screen media activity's effect on lack of perseverance, brain morphology, and cognitive performance, and propose new causal inference between genetics and screen time. Lastly, the algorithm used in this thesis project inferred reasonable causal orders and can be seen as a very good complement to today's causal modeling. / Skärmaktivitet som att titta på TV och video, spela videospel och använda sociala medier har blivit en populär fritidsaktivitet för barn och ungdomar. Effekten av skärmtid har varit ett mycket debatterat ämne; det finns dock fortfarande mycket lite kunskap om det. Med hjälp av datasetet från Adolescent Brain Cognitive Development långtidsstudien kunde 4 217 ungdomar, som uppfyllde specifika krav, väljas ut för detta avhandlingsprojekt efter bearbetning av datan. Detta avhandlingsprojekt undersökte kausal ordning mellan genetisk effekt (Polygenic scores (PGS) för kognitiv prestation), skärmtidsaktivitet, hjärnmorfologi (strukturell Magnet Resonans Imaging (sMRI) för hjärnans ytarea och hjärnbarks tjocklek), brist på ihärdighet och kognitiv förmåga (kristalliserad IQ) med en maskininlärningsalgoritm DirectLiNGAM. Tydlig korrelation mellan skärmaktivitet och PGS hittades för alla typer av skärmaktiviteter men endast videospel och sociala medier korrelerade till den globala ytarean. Dessutom verkar TV och video påverka brist på ihärdighet och brist på ihärdighet i sin tur påverkar hur mycket tid som spenderas på videospel. Dessa resultat antyder att olika typer av sociala medier inte är så lika som vi trodde och kan påverka ungdomar olika. Sammanlagt stöder dessa upptäckter tidigare forskning om skärmtidseffekt på brist på ihärdighet, hjärnmorfologi och kognitiv förmåga och föreslår en ny kausal inferens mellan genetik och skärmtid. Slutligen ledde algoritmen som användes i detta avhandlingsprojekt fram till rimliga kausala ordningar och kan ses som ett mycket bra komplement till dagens kausala modellering.

Master’s thesis is aimed at familiarization with the principles of measurement and data processing functional magnetic resonance, focusing on the analysis of effective connectivity using dynamic causal modelling (DCM). The practical part includes three main thematic areas relating to the description of the processing and evaluation of measured or simulated data. First, there is on sample dataset described the neuroscientific SPM toolbox to analyze measured data. Then follows introduction of the proposed approach with which is investigated the behavior of the model estimation neural interactions with respect to the change of input parameters. This phenomenon is also simulated and on base of achieved results is recommended optimal approach to analyzing effective connectivity using dynamic causal modeling for the group of subjects. The last circuit in the practical part is assessment of shift the coordinates of brain areas on dynamic causal modelling results for the group of subjects from the data obtained from real measurements. Obtained results from simulated data and the results obtained from measured data are evaluated and discussed in the final part.

The blood oxygen level dependent functional MRI (BOLD-fMRI) signal is an indirect measure of the neuronal activity that most BOLD studies are interested in. This thesis uses generative embedding algorithms to investigate some of the challenges and opportunities that this presents for BOLD imaging. It is standard practice to analyse BOLD signals using general linear models (GLMs) that assume fixed neurovascular coupling. However, this assumption may cause false positive or negative neural activations to be detected if the biological manifestations of brain diseases, disorders and pharmaceutical drugs (termed "neuromodulations") alter this coupling. Generative embedding can help overcome this problem by identifying when a neuromodulation confounds the standard GLM. When applied to anaesthetic neuromodulations found in preclinical imaging data, Fentanyl has the smallest confounding effect and Pentobarbital has the largest, causing extremely significant neural activations to go undetected. Half of the anaesthetics tested caused overestimation of the neuronal activity but the other half caused underestimation. The variability in biological action between anaesthetic modulations in identical brain regions of genetically similar animals highlights the complexity required to comprehensively account for factors confounding neurovascular coupling in GLMs generally. Generative embedding has the potential to augment established algorithms used to compensate for these variations in GLMs without complicating the standard (ANOVA) way of reporting BOLD results. Neuromodulation of neurovascular coupling can also present opportunities, such as improved diagnosis, monitoring and understanding of brain diseases accompanied by neurovascular uncoupling. Information theory is used to show that the discriminabilities of neurodegenerative-diseased and healthy generative posterior parameter spaces make generative embedding a viable tool for these commercial applications, boasting sensitivity to neurovascular coupling nonlinearities and biological interpretability. The value of hybrid neuroimaging systems over separate neuroimaging technologies is found to be greatest for early-stage neurodegenerative disease.

This PhD thesis investigated the role of the basal ganglia in memory and motor inhibition. Recent neuroimaging evidence suggests a supramodal network of inhibition involving the lateral prefrontal cortex. Here we examined whether this supramodal network also includes subcortical structures, such as the basal ganglia. Despite their well-established role in motor control, the basal ganglia are repeatedly activated but never interpreted during memory inhibition. We first used a series of meta-analyses to confirm the consistent involvement of the basal ganglia across studies using memory and motor inhibition tasks (including the Go/No-Go, Think/No-Think, and Stop-signal tasks), and discovered that there may be different subprocesses of inhibition. For instance, while the Go/No-Go task may require preventing a response from taking place, the Think/No-Think and Stop-signal tasks may require cancelling an emerging or ongoing response. We then conducted an fMRI study to examine how the basal ganglia interact with other putative supramodal regions (e.g., DLPFC) to achieve memory and motor inhibition during prevention and cancellation. Through dynamic causal modelling (DCM), we found that both DLPFC and basal ganglia play effective roles to achieve inhibition in the task-specific regions (hippocampus for memory inhibition; primary motor cortex (M1) for motor inhibition). Specifically, memory inhibition requires a DLPFC-basal ganglia-hippocampus pathway, whereas motor inhibition requires a basal ganglia-DLPFC-M1 pathway. We correlated DCM coupling parameters with behavioural indices to examine the relationship between network dynamics during prevention and cancellation and the successfulness of inhibition. However, due to constraints with DCM parameter estimates, caution is necessary when interpreting these results. Finally, we used diffusion weighted imaging to explore the anatomical connections supporting functions and behaviour. Unfortunately, we were unable to detect any white matter variability in relation to effective connectivity or behaviour during the prevention or cancellation processes of memory and motor inhibition at this stage. This PhD thesis provides essential INITIAL evidence that not only are the basal ganglia consistently involved in memory and motor inhibition, but these structures are effectively engaged in these tasks, achieving inhibition through task-specific pathways. We will discuss our findings, interpretations, and future directions in the relevant chapters.

This thesis deals with functional magnetic resonance imaging (fMRI), in particular with dynamic causal modelling (DCM) as one of the methods for effective brain connectivity analysis. It has been studied the effect of signal coordinates selection, which was used as an input of DCM analysis, on its results based on simulated data testing. For this purpose, a data simulator has been created and described in this thesis. Furthermore, the methodology of testing the influence of the coordinates selection on DCM results has been reported. The coordinates shift rate has been simulated by adding appropriate levels of various types of noise signals to the BOLD signal. Consequently, the data has been analyzed by DCM. The program has been supplemented by a graphical user interface. To determine behaviour of the model, Monte Carlo simulations have been applied. Results in the form of dependence of incorrectly estimated connections between brain areas on the level of the noise signals have been processed and discussed.

Erfolgreiche Unternehmen denken intensiv über den eigentlichen Nutzen ihres Unternehmens für Kunden nach. Diese versuchen, ihrer Konkurrenz voraus zu sein, und zwar durch gute Ideen, Innovationen und Kreativität. Dabei wird Erfolg anhand von Metriken gemessen, wie z.B. der Anzahl der loyalen Kunden oder der Anzahl der Käufer. Gegeben, dass der Wettbewerb durch die Globalisierung, Deregulierung und technologische Innovation in den letzten Jahren angewachsen ist, spielen die richtigen Entscheidungen für den Erfolg gerade im operativen Geschäft der sämtlichen Bereiche des Unternehmens eine zentrale Rolle. Vor diesem Hintergrund entstammen die in der vorliegenden Arbeit zur Evaluation der Methoden des maschinellen Lernens untersuchten Entscheidungsprobleme vornehmlich der Entscheidungsunterstützung. Hierzu gehören Klassifikationsprobleme wie die Kreditwürdigkeitsprüfung im Bereich Credit Scoring und die Effizienz der Marketing Campaigns im Bereich Direktmarketing. In diesem Kontext ergaben sich Fragestellungen für die korrelativen Modelle, nämlich die Untersuchung der Eignung der Verfahren des maschinellen Lernens für den Bereich des Credit Scoring, die Kalibrierung der Wahrscheinlichkeiten, welche mithilfe von Verfahren des maschinellen Lernens erzeugt werden sowie die Konzeption und Umsetzung einer Synergie-Heuristik zwischen den Methoden der klassischen Statistik und Verfahren des maschinellen Lernens. Desweiteren wurden kausale Modelle für den Bereich Direktmarketing (sog. Uplift-Effekte) angesprochen. Diese Themen wurden im Rahmen von breit angelegten empirischen Studien bearbeitet. Zusammenfassend ergibt sich, dass der Einsatz der untersuchten Verfahren beim derzeitigen Stand der Forschung zur Lösung praxisrelevanter Entscheidungsprobleme sowie spezifischer Fragestellungen, welche aus den besonderen Anforderungen der betrachteten Anwendungen abgeleitet wurden, einen wesentlichen Beitrag leistet. / Nowadays right decisions, being it strategic or operative, are important for every company, since these contribute directly to an overall success. This success can be measured based on quantitative metrics, for example, by the number of loyal customers or the number of incremental purchases. These decisions are typically made based on the historical data that relates to all functions of the company in general and to customers in particular. Thus, companies seek to analyze this data and apply obtained knowlegde in decision making. Classification problems represent an example of such decisions. Classification problems are best solved, when techniques of classical statistics and these of machine learning are applied, since both of them are able to analyze huge amount of data, to detect dependencies of the data patterns, and to produce probability, which represents the basis for the decision making. I apply these techniques and examine their suitability based on correlative models for decision making in credit scoring and further extend the work by causal predictive models for direct marketing. In detail, I analyze the suitability of techniques of machine learning for credit scoring alongside multiple dimensions, I examine the ability to produce calibrated probabilities and apply techniques to improve the probability estimations. I further develop and propose a synergy heuristic between the methods of classical statistics and techniques of machine learning to improve the prediction quality of the former, and finally apply conversion models to turn machine learning techqiques to account for causal relationship between marketing campaigns and customer behavior in direct marketing. The work has shown that the techniques of machine learning represent a suitable alternative to the methods of classical statistics for decision making and should be considered not only in research but also should find their practical application in real-world practices.