Custom coprocessor based matrix algorithms for image and signal processing

Matrix algorithms are important in many types of applications including image and signal processing. A close examination of the algorithms used in these, and related, applications reveals that many of the fundamental actions involve matrix algorithms such as matrix multiplication. This paper presents an investigation into the design and implementation of different matrix algorithms such as matrix operations, matrix transforms and matrix decompositions using a novel custom coprocessor system for MATrix algorithms based on Reconfigurable Computing (RCMAT). The proposed RCMAT architectures are scalable, modular and require less area and time complexity with reduced latency when compared with existing structures.

The sites of neural adaptation induced by resistance training in humans

Although it has long been supposed that resistance training causes adaptive changes in the CNS, the sites and nature of these adaptations have not previously been identified. In order to determine whether the neural adaptations to resistance training occur to a greater extent at cortical or subcortical sites in the CNS, we compared the effects of resistance training on the electromyographic (EMG) responses to transcranial magnetic (TMS) and electrical (TES) stimulation. Motor evoked potentials (MEPs) were recorded from the first dorsal interosseous muscle of 16 individuals before and after 4 weeks of resistance training for the index finger abductors (n=8), or training involving finger abduction-adduction without external resistance (n=8). TMS was delivered at rest at intensities from 5% below the passive threshold to the maximal output of the stimulator. TMS and TES were also delivered at the active threshold intensity while the participants exerted torques ranging from 5 to 60% of their maximum voluntary contraction (MVC) torque. The average latency of MEPs elicited by TES was significantly shorter than that of TMS MEPs (TES latency=21.5+/-1.4 ms; TMS latency=23.4+/-1.4 ms; P

Online CORDIC algorithm and VLSI architecture for implementing QR-array processors

A novel most significant digit first CORDIC architecture is presented that is suitable for the VLSI design of systolic array processor cells for performing QR decomposition. This is based on an on-line CORDIC algorithm with a constant scale factor and a latency independent of the wordlength. This has been derived through the extension of previously published CORDIC algorithms. It is shown that simplifying the calculation of convergence bounds also greatly simplifies the derivation of suitable VLSI architectures. Design studies, based on a 0.35-µ CMOS standard cell process, indicate that 20 such QR processor cells operating at rates suitable for radar beamfoming can be readily accommodated on a single chip.

Responses of calves to acute stress: Individual consistency and relations between behavioral and physiological measures

The present study examined the consistency over time of individual differences in behavioral and physiological responsiveness of calves to intuitively alarming test situations as well as the relationships between behavioral and physiological measures. Twenty Holstein Friesian heifer calves were individually subjected to the same series of two behavioral and two hypothalamo-pituitary-adrenocortical (HPA) axis reactivity tests at 3, 13 and 26 weeks of age. Novel environment (open field, OF) and novel object (NO) tests involved measurement of behavioral, plasma cortisol and heart rate responses. Plasma ACTH and/or cortisol response profiles were determined after administration of exogenous CRH and ACTH, respectively, in the HPA axis reactivity tests. Principal component analysis (PCA) was used to condense correlated measures within ages into principal components reflecting independent dimensions underlying the calves' reactivity. Cortisol responses to the OF and NO tests were positively associated with the latency to contact and negatively related to the time spent in contact with the NO. Individual differences in scores of a principal component summarizing this pattern of inter-correlations, as well as differences in separate measures of adrenocortical and behavioral reactivity in the OF and NO tests proved highly consistent over time. The cardiac response to confinement in a start box prior to the OF test was positively associated with the cortisol responses to the OF and NO tests at 26 weeks of age. HPA axis reactivity to ACTH or CRH was unrelated to adrenocortical and behavioral responses to novelty. These findings strongly suggest that the responsiveness of calves was mediated by stable individual characteristics. Correlated adrenocortical and behavioral responses to novelty may reflect underlying fearfulness, defining the individual's susceptibility to the elicitation of fear. Other independent characteristics mediating reactivity may include activity or coping style (related to locomotion) and underlying sociality (associated with vocalization). (c) 2005 Elsevier Inc. All rights reserved.

Synchrophasor Broadcast over Internet Protocol for Distributed Generator Synchronization

Synchronous islanded operation involves continuously holding an islanded power network in virtual synchronism with the main power system to aid paralleling and avoid potentially damaging out-of-synchronism reclosure. This requires phase control of the generators in the island and the transmission of a reference signal from a secure location on the main power system. Global positioning system (GPS) time-synchronized phasor measurements transmitted via an Internet protocol (IP) are used for the reference signal. However, while offering low cost and a readily available solution for distribution networks, IP communications have variable latency and are susceptible to packet loss, which can make time-critical control applications difficult. This paper investigates the ability of the phase-control system to tolerate communications latency. Phasor measurement conditioning algorithms that can tolerate latency are used in the phase-control loop of a 50-kVA diesel generator. © 2010 IEEE.

Pim2 cooperates with PML-RAR alpha to induce acute myeloid leukemia in a bone marrow transplantation model

Although the potential role of Pim2 as a cooperative oncogene has been well described in lymphoma, its role in leukemia has remained largely unexplored. Here we show that high expression of Pim2 is observed in patients with acute promyelocytic leukemia (APL). To further characterize the cooperative role of Pim2 with promyelocytic leukemia/retinoic acid receptor alpha (PML/RAR alpha), we used a well-established PML-RAR alpha (PR alpha) mouse model. Pim2 coexpression in PR alpha-positive hematopoietic progenitor cells (HPCs) induces leukemia in recipient mice after a short latency. Pim2-PR alpha cells were able to repopulate mice in serial transplantations and to induce disease in all recipients. Neither Pim2 nor PR alpha alone was sufficient to induce leukemia upon transplantation in this model. The disease induced by Pim2 overexpression in PR alpha cells contained a slightly higher fraction of immature myeloid cells, compared with the previously described APL disease induced by PR alpha. However, it also clearly resembled an APL-like phenotype and showed signs of differentiation upon all-trans retinoic acid (ATRA) treatment in vitro. These results support the hypothesis that Pim2, which is also a known target of Flt3-ITD (another gene that cooperates with PML-RAR alpha), cooperates with PR alpha to induce APL-like disease. (Blood. 2010; 115(22): 4507-4516)

Design of quantum-dot cellular automata circuits using cut-set retiming

As a potential alternative to CMOS technology, QCA provides an interesting paradigm in both communication and computation. However, QCAs unique four-phase clocking scheme and timing constraints present serious timing issues for interconnection and feedback. In this work, a cut-set retiming design procedure is proposed to resolve these QCA timing issues. The proposed design procedure can accommodate QCAs unique characteristics by performing delay-transfer and time-scaling to reallocate the existing delays so as to achieve efficient clocking zone assignment. Cut-set retiming makes it possible to effectively design relatively complex QCA circuits that include feedback. It utilizes the similar characteristics of synchronization, deep pipelines and local interconnections common to both QCA and systolic architectures. As a case study, a systolic Montgomery modular multiplier is designed to illustrate the procedure. Furthermore, a nonsystolic architecture, an S27 benchmark circuit, is designed and compared with previous designs. The comparison shows that the cut-set retiming method achieves a more efficient design, with a reduction of 22%, 44%, and 46% in terms of cell count, area, and latency, respectively.

Auditory sensory deficits in developmental dyslexia: A longitudinal ERP study

The core difficulty in developmental dyslexia across languages is a "phonological deficit", a specific difficulty with the neural representation of the sound structure of words. Recent data across languages suggest that this phonological deficit arises in part from inefficient auditory processing of the rate of change of the amplitude envelope at syllable onset (inefficient sensory processing of rise time). Rise time is a complex percept that also involves changes in duration and perceived intensity. Understanding the neural mechanisms that give rise to the phonological deficit in dyslexia is important for optimising educational interventions. In a three-deviant passive 'oddball' paradigm and a corresponding blocked 'deviant-alone' control condition we recorded ERPs to tones varying in rise time, duration and intensity in children with dyslexia and typically developing children longitudinally. We report here results from test Phases 1 and 2, when participants were aged 8-10. years. We found an MMN to duration, but not to rise time nor intensity deviants, at both time points for both groups. For rise time, duration and intensity we found group effects in both the Oddball and Blocked conditions. There was a slower fronto-central P1 response in the dyslexic group compared to controls. The amplitude of the P1 fronto-centrally to tones with slower rise times and lower intensity was smaller compared to tones with sharper rise times and higher intensity in the Oddball condition, for children with dyslexia only. The latency of this ERP component for all three stimuli was shorter on the right compared to the left hemisphere, only for the dyslexic group in the Blocked condition. Furthermore, we found decreased N1c amplitude to tones with slower rise times compared to tones with sharper rise times for children with dyslexia, only in the Oddball condition. Several other effects of stimulus type, age and laterality were also observed. Our data suggest that neuronal responses underlying some aspects of auditory sensory processing may be impaired in dyslexia. © 2011 Elsevier Inc.

Cache-Integrated Network Interfaces: Flexible On-Chip Communication and Synchronization for Large-Scale CMPs

Per-core scratchpad memories (or local stores) allow direct inter-core communication, with latency and energy advantages over coherent cache-based communication, especially as CMP architectures become more distributed. We have designed cache-integrated network interfaces, appropriate for scalable multicores, that combine the best of two worlds – the flexibility of caches and the efficiency of scratchpad memories: on-chip SRAM is configurably shared among caching, scratchpad, and virtualized network interface (NI) functions. This paper presents our architecture, which provides local and remote scratchpad access, to either individual words or multiword blocks through RDMA copy. Furthermore, we introduce event responses, as a technique that enables software configurable communication and synchronization primitives. We present three event response mechanisms that expose NI functionality to software, for multiword transfer initiation, completion notifications for software selected sets of arbitrary size transfers, and multi-party synchronization queues. We implemented these mechanisms in a four-core FPGA prototype, and measure the logic overhead over a cache-only design for basic NI functionality to be less than 20%. We also evaluate the on-chip communication performance on the prototype, as well as the performance of synchronization functions with simulation of CMPs with up to 128 cores. We demonstrate efficient synchronization, low-overhead communication, and amortized-overhead bulk transfers, which allow parallelization gains for fine-grain tasks, and efficient exploitation of the hardware bandwidth.

A Low Complexity Real-time MIMO-Preprocessing For Fixed Complexity Sphere Decoder

Modern Multiple-Input Multiple-Output (MIMO) communication systems place huge demands on embedded processing resources in terms of throughput, latency and resource utilization. State-of-the-art MIMO detector algorithms, such as Fixed-Complexity Sphere Decoding (FSD), rely on efficient channel preprocessing involving numerous calculations of the pseudo-inverse of the channel matrix by QR Decomposition (QRD) and ordering. These highly complicated operations can quickly become the critical prerequisite for real-time MIMO detection, exaggerated as the number of antennas in a MIMO detector increases. This paper describes a sorted QR decomposition (SQRD) algorithm extended for FSD, which significantly reduces the complexity and latency
of this preprocessing step and increases the throughput of MIMO detection. It merges the calculations of the QRD and ordering operations to avoid multiple iterations of QRD. Specifically, it shows that SQRD reduces the computational complexity by over 60-70% when compared to conventional
MIMO preprocessing algorithms. In 4x4 to 7x7 MIMO cases, the approach suffers merely 0.16-0.2 dB reduction in Bit Error Rate (BER) performance.

Managing SMT Resource Usage through Speculative Instruction Window Weighting

Simultaneous multithreading processors dynamically share processor resources between multiple threads. In general, shared SMT resources may be managed explicitly, for instance, by dynamically setting queue occupation bounds for each thread as in the DCRA and Hill-Climbing policies. Alternatively, resources may be managed implicitly; that is, resource usage is controlled by placing the desired instruction mix in the resources. In this case, the main resource management tool is the instruction fetch policy which must predict the behavior of each thread (branch mispredictions, long-latency loads, etc.) as it fetches instructions.

On generating set index functions for randomized caches

Caches hide the growing latency of accesses to the main memory from the processor by storing the most recently used data on-chip. To limit the search time through the caches, they are organized in a direct mapped or set-associative way. Such an organization introduces many conflict misses that hamper performance. This paper studies randomizing set index functions, a technique to place the data in the cache in such a way that conflict misses are avoided. The performance of such a randomized cache strongly depends on the randomization function. This paper discusses a methodology to generate randomization functions that perform well over a broad range of benchmarks. The methodology uses profiling information to predict the conflict miss rate of randomization functions. Then, using this information, a search algorithm finds the best randomization function. Due to implementation issues, it is preferable to use a randomization function that is extremely simple and can be evaluated in little time. For these reasons, we use randomization functions where each randomized address bit is computed as the XOR of a subset of the original address bits. These functions are chosen such that they operate on as few address bits as possible and have few inputs to each XOR. This paper shows that to index a 2(m)-set cache, it suffices to randomize m+2 or m+3 address bits and to limit the number of inputs to each XOR to 2 bits to obtain the full potential of randomization. Furthermore, it is shown that the randomization function that we generate for one set of benchmarks also works well for an entirely different set of benchmarks. Using the described methodology, it is possible to reduce the implementation cost of randomization functions with only an insignificant loss in conflict reduction.

Primary motor cortex involvement in initial learning during visuomotor adaptation

Human motor behaviour is continually modified on the basis of errors between desired and actual movement outcomes. It is emerging that the role played by the primary motor cortex (M1) in this process is contingent upon a variety of factors, including the nature of the task being performed, and the stage of learning. Here we used repetitive TMS to test the hypothesis that M1 is intimately involved in the initial phase of sensorimotor adaptation. Inhibitory theta burst stimulation was applied to M1 prior to a task requiring modification of torques generated about the elbow/forearm complex in response to rotations of a visual feedback display. Participants were first exposed to a 30° clockwise (CW) rotation (Block A), then a 60° counterclockwise rotation (Block B), followed immediately by a second block of 30° CW rotation (A2). In the STIM condition, participants received 20s of continuous theta burst stimulation (cTBS) prior to the initial A Block. In the conventional (CON) condition, no stimulation was applied. The overt characteristics of performance in the two conditions were essentially equivalent with respect to the errors exhibited upon exposure to a new variant of the task. There were however, profound differences between the conditions in the latency of response preparation, and the excitability of corticospinal projections from M1, which accompanied phases of de-adaptation and re-adaptation (during Blocks B and A2). Upon subsequent exposure to the A rotation 24h later, the rate of re-adaptation was lower in the stimulation condition than that present in the conventional condition. These results support the assertion that primary motor cortex assumes a key role in a network that mediates adaptation to visuomotor perturbation, and emphasise that it is engaged functionally during the early phase of learning.

Training-induced modifications of corticospinal reactivity in severely affected stroke survivors

When permitted access to the appropriate forms of rehabilitation, many severely affected stroke survivors demonstrate a capacity for upper limb functional recovery well in excess of that formerly considered possible. Yet, the mechanisms through which improvements in arm function occur in such profoundly impaired individuals remain poorly understood. An exploratory study was undertaken to investigate the capacity for brain plasticity and functional adaptation, in response to 12-h training of reaching using the SMART Arm device, in a group of severely affected stroke survivors with chronic upper limb paresis. Twenty-eight stroke survivors were enroled. Eleven healthy adults provided normative data. To assess the integrity of ipsilateral and contralateral corticospinal pathways, transcranial magnetic stimulation was applied to evoke responses in triceps brachii during an elbow extension task. When present, contralateral motor-evoked potentials (MEPs) were delayed and reduced in amplitude compared to those obtained in healthy adults. Following training, contralateral responses were more prevalent and their average onset latency was reduced. There were no reliable changes in ipsilateral MEPs. Stroke survivors who exhibited contralateral MEPs prior to training achieved higher levels of arm function and exhibited greater improvements in performance than those who did not initially exhibit contralateral responses. Furthermore, decreases in the onset latency of contralateral MEPs were positively related to improvements in arm function. Our findings demonstrate that when severely impaired stroke survivors are provided with an appropriate rehabilitation modality, modifications of corticospinal reactivity occur in association with sustained improvements in upper limb function.

Aneurysms increase the risk of rebleeding after stereotactic radiosurgery for hemorrhagic arteriovenous malformations

BACKGROUND AND PURPOSE:
The purpose of this study was to define the risk of rebleeding after stereotactic radiosurgery (SRS) for hemorrhagic arteriovenous malformations with or without associated intracranial aneurysms.

METHODS:
Between 1987 and 2006, we performed Gamma Knife SRS on 996 patients with brain arteriovenous malformations; 407 patients had sustained an arteriovenous malformation hemorrhage. Sixty-four patients (16%) underwent prior embolization and 84 (21%) underwent prior surgical resection. The median target volume was 2.3 mL (range, 0.1-20.7 mL). The median margin dose was 20 Gy (range, 13.5-27 Gy).

RESULTS:
The overall rate of total obliteration defined by angiography or MRI was 56%, 77%, 80%, and 82% at 3, 4, 5, and 10 years, respectively. Before obliteration, 33 patients (8%) sustained an additional hemorrhage after SRS. The overall annual hemorrhage rate until obliteration after SRS was 1.3%. The presence of a patent aneurysm was significantly associated with an increased rehemorrhage risk after SRS (annual hemorrhage rate, 6.4%) compared with patients with a clipped or embolized aneurysm (annual hemorrhage rate, 0.8%; P=0.033).

CONCLUSIONS:
When an aneurysm is identified in patients with arteriovenous malformations selected for SRS, additional endovascular or surgical strategies should be considered to reduce the risk of bleeding during the latency interval.