The design and synthesis of stable organic radicals with delocalized spin density distribution and low energy optical and redox processes is central to the development of magneto-conducting materials. Towards this end, a generalized synthetic methodology has been developed allowing for the synthesis of a series of annelated benzonitronyl nitroxide (BNN) radicals. The BNN radicals exhibited remarkably low reduction potentials (~0.0 V vs SCE) and a near-infrared absorption attributed to a HOMO–SOMO charge-transfer excitation. The annelated BNN radicals were found to be less stable than the closely related tetramethyl nitronyl nitroxide radicals, particularly in solution. A series of π-delocalized and heteroaromatic radicals were synthesized to determine if the instability was due to the delocalization of electron density onto the carbon skeleton or the low reduction potential. DFT calculations with the EPR-II basis gave rise to calculated electronic structures that were consistent with EPR spectroscopy and suggested changes in spin density distribution are occurring with perturbation of the annelated ring. Cyclic voltammetry revealed the heteroaromatic and π-delocalized radicals had reduction potentials lower than BNN, with some systems reducing at potentials of 0.2 V vs SCE, comparable to that of 7,7,8,8-tetracyanoquinodimethane. The distribution of spin throughout the molecular framework and the low reduction potential of the annelated nitronyl nitroxide radicals were both found to contribute to the lowered stability of the annelated nitronyl nitroxides relative to the far less redox active tetramethyl nitronyl nitroxides. The low reduction potential of the BNN radicals suggested that incorporation into acceptor–donor (A–D) systems would allow for investigation of the role of charge transfer interactions on the electronic structure and properties of neutral open-shell A–D radicals. Two D–A–D radicals were prepared using metal catalyzed coupling and furoxan condensation methodologies which resulted in incorporation of a second donor in the C5 position of the BNN moiety. The radical D1–A–D2 triads, where D1 = thiophene and D2 = thiophene or phenyl, exhibited three intramolecular charge-transfer excitations (λmax = 550, 580 and 1000 nm) that were investigated by variable temperature absorption spectroscopy. Structural characterization of the triads in the solid state by single crystal and powder X-ray diffraction revealed slipped π stacks that arise from intermolecular π– π and D–A interactions, providing pathways for antiferromagnetic (AFM) and ferromagnetic (FM) exchange. While the phenyl substituted triad (Th–BNN–Ph) exhibited antiferromagnetic interactions and a room temperature conductivity of σRT = 10−7 S cm−1, the thienyl substituted derivative (Th–BNN–Th) exhibited short-range FM interactions and increased conductivity (σRT = 10−5 S cm−1), giving rise to an organic semiconductor exhibiting FM exchange. The differences in conductivity and magnetic behavior were rationalized by the degree of slippage dictated by an interplay between π– π and intermolecular D−A interactions. Finally, a series of BNN–D radicals were investigated where the donor ability of D was systematically varied from Eox = 2.30 V vs SCE (benzene) to 0.32 V vs SCE (tetrathiafulvalene). Calculations of the near-infrared charge transfer excitation suggested that the HOMO–SOMO gap could be significantly decreased with increasing donor ability, consistent with charge transfer theory. A subset of the series of BNN–D radicals with D = anisole, benzo[b]thiophene, N-methylindole, N-ethylcarbazole, and N,Ndiphenylaniline were synthesized. Solution state spectroscopic studies of the series by EPR and electronic absorption spectroscopy revealed spin-delocalized structures with extremely low reduction potentials (~0 V vs SCE). The solid state properties of the BNN–D radicals were investigated by magnetometry and room temperature conductivity measurements. Due to primarily steric interactions, weak D–A coupling was observed, with weak intermolecular interactions in the solid state leading to paramagnetic and insulating behaviour. The BNN-(N,N-diphenylaniline) radical structure was characterized by single crystal XRD and found to exist as well isolated radical moieties with extremely weak intermolecular interactions, consistent with magnetometry and conductivity measurements. / Graduate / 0490 / 0794

This thesis work presents an unique architecture with a capacity of 1.57-Mb storage including in-memory compuitng capability, leveraging state-of-the-art gate voltage assisted spin-orbit torque (VGSOT) magnetic random-access memory (MRAM) technology. Beyond its role as a non-volatile storage solution, this architecture facilitates a diverse array of In-Memory Computing (IMC) operations, inclusive of logic-inside-memory (LinM/LiM), in-memory-dot- product multiplication tailored for binary-neural-networks, and content-accessable memory (CAM). Our designed bit-cell proposed in this architecture occupies a compact area of 0.195 μm2 and exhibits remarkable performance metrics. It achieves impressive writing speeds of 200 MHz and reading speeds of 1.5 GHz, applicable to non-volatile storage tasks and LinM operations. Notably, the LinM functionality supports a wide range of logical operations such as AND, NAND, OR, NOR, and MAJ, while the CAM feature enables efficient data searches of up to 1024 bits. Furthermore, in performance evaluations conducted using the MNIST and FMNIST datasets with a BNN model structured as 512-512-10 (input layer - hidden layer - output layer), the proposed VGSOT MRAM demonstrates exceptional inference accuracy. Specifically, it achieves a high accuracy rate of 97.40% for the MNIST dataset and 84.15% for the FMNIST dataset. In comparison to the 2T1R SOT-MRAM technology, the proposed VGSOT MRAM showcases significant advancements in read performance and reliability metrics. Notably, it features a 65.74% reduction in bit-cell area, alongside 84.78% and 33.4% lower read-write power consumption and 54.11% and 30.57% reduced LinM power consumption, respectively. / Master of Science / This work brings forth towards a new technology called VGSOT MRAM, which is a type of memory device that can store information without using extensive power. Its part of a larger architecture called IMC, which has many useful features. One of the main advantages of this technology is that it can perform different operations while storing data. For example, it can do calculations, search for specific information, and perform tasks for artificial intelligence networks. The design of the memory cells is also very efficient, taking up a small amount of space. In terms of performance, this technology is quite impressive. It can write data very quickly, at a speed of 250 million times per second, and read data even faster, at 1.67 billion times per second. It can also perform different logical operations, like AND, OR, and NAND, which are important for many computing tasks when tested with real-world tasks, such as recognizing images, this technology showed excellent accuracy. It achieved a recognition accuracy of 97.40% for the MNIST dataset and 84.15% for the FMNIST dataset, which is quite good. Compared to other similar technologies, this VGSOT MRAM has some advantages. It takes up less space, uses less power when reading and writing data, and consumes less power when performing calculations. These improvements make it a promising option for future devices.

Οι μεταλλαγμένοι μύες weaver αποτελούν ένα μοναδικό γενετικό μοντέλο της νόσου Πάρκινσον. Πρόκειται για μια νευροβιολογική μετάλλαξη η οποία συμβολίζεται ως wv και περιγράφηκε για πρώτη φορά από τον Lane (1964). Πρόκειται για μετάλλαξη με λάθος νόημα στο γονίδιο Girk 2 που βρίσκεται στο χρωμόσωμα 16 του ποντικού σε μια περιοχή ανάλογη με την περιοχή του 21 χρωμοσώματος του ανθρώπου που ευθύνεται για το σύνδρομο Down και οδηγεί σε τροποποίηση της λειτουργίας ενός διαύλου καλίου. Επειδή ο εγκέφαλος των θηλαστικών είναι από τους πιο ευαίσθητους ιστούς πολλές μελέτες έχουν γίνει για τις επιδράσεις των φαρμάκων (νευροστεροειδών), στην παθογένεια και στην αντιμετώπιση των νευροεκφυλιστικών νόσων. Ο παράγοντας BNN 50 είναι ένα νευροστεροειδές που έχει συνώνυμο το BNN 20. Παράγεται από το ανδρογόνο νευροστεροειδές δεϋδροεπιανδροστερόνη (DHEA) με εποξείδωση της 17-κετομάδας του με χρήση trimethylsulfonium methylide (τριμεθυλοσούλφωνο-μεθυλίδιο). Έχει το πλεονέκτημα ότι δεν μεταβολίζεται σε οιστρογόνα ενδογενώς, όπως το DHEA, κι επομένως ενδέχεται να είναι καταλληλότερο για φαρμακευτική χρήση. Σκοπός της παρούσας μελέτης είναι να διερευνηθεί η επίδραση στο οξειδωτικό στρες του νευροστεροειδούς BNN 50 σε εγκεφάλους μυών κατά την 21 ημέρα με ταυτόχρονο προσδιορισμό των επιπέδων υπεροξείδωσης λιπιδίων και πρωτεϊνών στις ακόλουθες περιοχές: α) CE (παρεγκεφαλίδα), β) HIP (ιππόκαμπος), γ) CX (εγκεφαλικός φλοιός). Για το σκοπό αυτό τα πειραματόζωα χωρίστηκαν στις ακόλουθες 4 ομάδες: 1) φυσιολογικοί μύες (control) χωρίς BNN 50, με saline 2) control με BNN 50, 3) weaver χωρίς BNN 50 με saline και 4) weaver με ΒΝΝ 50. Οι μύες (control-weaver) δέχονταν BNN 50 (100 mg/kg σωματικού βάρους) ημερησίως από την πρώτη μέρα της ζωής (με ενδοπεριτοναϊκή ένεση) και για διάστημα 18 ημερών. Μετά την τελευταία ένεση, οι μύες θανατώθηκαν και οι εγκεφαλικοί ιστοί απομονώθηκαν σε ειδικό αντιοξειδωτικό ρυθμιστικό διάλυμα. Για τον προσδιορισμό του βαθμού λιπιδικής υπεροξείδωσης στα ομογενοποιήματα των ιστών χρησιμοποιήθηκε ως δείκτης η MDA (μηλονική δυαλδεύδη), σε ελεύθερη μορφή και συνδεδεμένη με πρωτεΐνες (FrMDA και PrMDA, αντιστοίχως). Επίσης, ποσοτικοποιήθηκαν και τα υδροϋπεροξείδια των λιπιδίων και των πρωτεϊνών (LOOH και PrOOH, αντιστοίχως). Όσον αφορά τα αποτελέσματα βρέθηκε ότι: (1) το νευροστεροειδές BNN 50 δρα αντιοξειδωτικά στις δυο από τις τρεις περιοχές εγκεφάλου που μελετήθηκαν (ιππόκαμπος και παρεγκεφαλίδα), και (2) οι μύες weaver στους οποίους είχε χορηγηθεί το φάρμακο είχαν χαμηλότερες τιμές σε τρεις από τους τέσσερις δείκτες ελέγχου για λιπιδική και πρωτεϊνική υπεροξείδωση εκτός από τα πρωτεϊνικά υπεροξείδια. Συνεπώς η αντιοξειδωτική δράση των νευροστεροειδών ενδέχεται να συντελέσει στην ανάπτυξη νέων μεθόδων θεραπείας για την αντιμετώπιση των νευροεκφυλιστικών νόσων. / The mutant mouse weaver constitutes a unique genetic model of Parkinson's disease. It is a neurobiological mutation denoted wv and first described by Lane (1964). This mutation in the wrong sense Girk 2 gene located on chromosome 16 of the mouse in a region similar to the region of human chromosome 21 that is responsible for the syndrome Down and leads to modulation of the function of a potassium channel. Because the brain of mammals is among the most sensitive tissues many studies have been done on the effects of drugs (neurosteroids) in the pathogenesis and treatment of neurodegenerative diseases. The agent BNN 50 is a neuroactive steroids having synonymous BNN 20. Produced by neuroactive steroids androgen dehydroepiandrosterone (DHEA) by epoxidation of 17- keto group of using trimethylsulfonium methylide (trimethylsulfonium - methylide). It has the advantage that it is metabolized to endogenous estrogens, such as DHEA, and therefore may be suitable for pharmaceutical use. The purpose of this study is to investigate the effect of oxidative stress neurosteroid BNN 50 in brains muscle at day 21 with simultaneous determination of the levels of lipid peroxidation and protein in the following areas: a) CE (cerebellum), b) HIP (hippocampus) c) CX (cerebral cortex). For this purpose, the animals were divided in the following four groups: 1) normal muscles (control) without BNN 50, with saline 2) control with BNN 50, 3) weaver without BNN 50 with saline and 4) weaver with BNN 50. Muscles accept BNN 50 (100 mg / kg body weight) per day from the first day of life (by intraperitoneal injection) and for 18 days. After the last injection, the mice were sacrificed and brain tissues were isolated in a special antioxidant buffer. For determining the degree of lipid peroxidation in tissue homogenates was used as the index of MDA (dyaldefdi malonate), in free form and bound protein (FrMDA and PrMDA, respectively). Also quantitated and hydroperoxides of lipids and proteins (LOOH and PrOOH, respectively). Regarding the results found that: (1) the neuroactive steroids BNN 50 acts antioxidants in two of the three brain regions studied (hippocampus and cerebellum), and (2) the weaver mice dosed with the drug were lower in three of four indicators for monitoring lipid peroxidation and protein than the protein peroxides. Therefore, the antioxidant activity of neurosteroids may contribute to the development of new therapies for the treatment of neurodegenerative diseases.

Image classification is a memory- and compute-intensive task. It is difficult to implement high-speed image classification algorithms on resource-limited systems like FPGAs and embedded computers. Most image classification algorithms require many fixed- and/or floating-point operations and values. In this work, we explore the use of binary values to reduce the memory and compute requirements of image classification algorithms. Our objective was to implement these algorithms on resource-limited systems while maintaining comparable accuracy and high speeds. By implementing high-speed image classification algorithms on resource-limited systems like embedded computers, FPGAs, and ASICs, automated visual inspection can be performed on small low-powered systems. Industries like manufacturing, medicine, and agriculture can benefit from compact, high-speed, low-power visual inspection systems. Tasks like defect detection in manufactured products and quality sorting of harvested produce can be performed cheaper and more quickly. In this work, we present ECO Jet Features, an algorithm adapted to use binary values for visual inspection. The ECO Jet Features algorithm ran 3.7x faster than the original ECO Features algorithm on embedded computers. It also allowed the algorithm to be implemented on an FPGA, achieving 78x speedup over full-sized desktop systems, using a fraction of the power and space. We reviewed Binarized Neural Nets (BNNs), neural networks that use binary values for weights and activations. These networks are particularly well suited for FPGA implementation and we compared and contrasted various FPGA implementations found throughout the literature. Finally, we combined the deep learning methods used in BNNs with the efficiency of Jet Features to make Neural Jet Features. Neural Jet Features are binarized convolutional layers that are learned through deep learning and learn classic computer vision kernels like the Gaussian and Sobel kernels. These kernels are efficiently computed as a group and their outputs can be reused when forming output channels. They performed just as well as BNN convolutions on visual inspection tasks and are more stable when trained on small models.

Η νόσος του Parkinson χαρακτηρίζεται από τη βαθμιαία, εκλεκτική νευροεκφύλιση των ντοπαμινεργικών νευρώνων της μελαινοραβδωτής οδού. Η μειωμένη δραστηριοποίηση των ντοπαμινεργικών υποδοχέων που προκαλείται από την ανεπάρκεια ντοπαμίνης τροποποιεί τη λειτουργία των βασικών γαγγλίων και αναστέλλει τα κινητικά συστήματα. Ιδανικό πειραματικό μοντέλο αποτελεί το μοντέλο weaver, το οποίο εμφανίζει το ίδιο μοτίβο νευροεκφύλισης με τους παρκινσονικούς ασθενείς καθώς και περισσότερη από 70% μείωση της ντοπαμίνης στο ραβδωτό σώμα. Το γεγονός πως δεν υπάρχει σήμερα αποτελεσματική θεραπεία που να σταματά ή να αναστρέφει την εξέλιξη της νόσου, δημιουργεί την ανάγκη ανακάλυψης ενός φαρμακευτικού σχήματος το οποίο θα έχει νευροπροστατευτική δράση και θα περιορίζει τις παρενέργειες. Προηγούμενα αποτελέσματα της ομάδας μας δείχνουν σημαντική επιβίωση των ντοπαμινεργικών κυττάρων στο μοντέλο weaver μετά από χρόνια χορήγηση του ενδογενούς νευροστεροειδούς DHEAS, του χημικού αναλόγου του ΒΝΝ-50 (το οποίο δε μεταβολίζεται σε οιστρογόνα) και του φαρμακευτικού συνδυασμού του ΒΝΝ-50 με τη Ν-ακετυλοκυστεΐνη (NAC), με το τελευταίο να επαναφέρει πλήρως τον αριθμό των κυττάρων στη μέλαινα ουσία. Στην παρούσα εργασία θελήσαμε να διερευνήσουμε τους μηχανισμούς μέσω των οποίων επιτυγχάνεται η νευροπροστασία που προκαλεί η συγχορήγηση του συνδυασμού BNN-50/ΝAC. Για το σκοπό αυτό, αξιολογήσαμε την αντιαποπτωτική, καθώς και την αντιοξειδωτική δράση του σχήματος BNN-50/NAC. Οι δύο δείκτες επιλέχθηκαν λαμβάνοντας υπόψιν τον κεντρικό ρόλο του αποπτωτικού θανάτου στη διαδικασία της νευροεκφύλισης, καθώς και το ρόλο του οξειδωτικού στρες στην παθογένεια της νόσου. Τα αποτελέσματα μας υποδεικνύουν πλειοτροπική δράση του φαρμακευτικού συνδυασμού BNN-50/NAC, η οποία εκφράζεται μέσω της ισχυρής αντιαποπτωτικής και αντιοξειδωτικής του δράσης. / Parkinson 's disease is characterized by the progressive, selective neurodegeneration of the dopaminergic neurons of the nigrostriatal pathway. The decreased activation of dopamine receptors caused by insufficient dopamine levels, modifies the function of the basal ganglia circuit and inhibits the mobility systems. The weaver model consists an ideal model for neuroprotection studies, which exhibits the same pattern of neurodegeneration as the parkinsonian patients and more than 70% decrease of dopamine in the striatum. The fact that there is currently no effective treatment to attenuate or reverse the disease progression, creates the need for discovery of a drug combination which will exhibit neuroprotective effect and reduce the side effects. Previous results of our group, have shown a significant survival of dopaminergic neurons in weaver mice after chronic administration of endogenous neurosteroid DHEA-S, the chemical analog BNN-50 (which is not metabolized to estrogens) and the combination of the BNN-50 with N-acetylcysteine (NAC,with the latter combination completely rescuing the number of dopaminergic cells of the substantia nigra. The aim of this study was to investigate the mechanisms of neuroprotection induced by coadministration of combination BNN-50/NAC. For this purpose, we evaluated the possible antiapoptotic and antioxidant action of the BNN-50/NAC combination. Our results suggest a pleiotropic effect of the BNN-50/NAC drug combination, that is expressed through strong antiapoptotic and antioxidant activity.

Artificial neural networks (ANNs) are used in many types of computing applications. Traditionally, ANNs have been implemented in software, executing on CPUs and even GPUs, which capitalize on the parallelizable nature of ANNs. More recently, FPGAs have become a target platform for ANN implementations due to their relatively low cost, low power, and flexibility. Some safety-critical applications could benefit from ANNs, but these applications require a certain level of reliability. SRAM-based FPGAs are sensitive to single-event upsets (SEUs), which can lead to faults and errors in execution. However there are techniques that can mask such SEUs and thereby improve the overall design reliability. This thesis evaluates the SEU reliability of neural networks implemented in SRAM-based FPGAs and investigates mitigation techniques against upsets for two case studies. The first was based on the LeNet-5 convolutional neural network and was used to test an implementation with both fault injection and neutron radiation experiments, demonstrating that our fault injection experiments could accurately evaluate SEU reliability of the networks. SEU reliability was improved by selectively applying TMR to the most critical layers of the design, achieving a 35% improvement reliability at an increase in 6.6% resources. The second was an existing neural network called BNN-PYNQ. While the base design was more sensitive to upsets than the CNN previous tested, the TMR technique improved the reliability by approximately 7× in fault injection experiments.