Brain-inspired computing needs a master plan.

New computing technologies inspired by the brain promise fundamentally different ways to process information with extreme energy efficiency and the ability to handle the avalanche of unstructured and noisy data that we are generating at an ever-increasing rate. To realize this promise requires a brave and coordinated plan to bring together disparate research communities and to provide them with the funding, focus and support needed. We have done this in the past with digital technologies; we are in the process of doing it with quantum technologies; can we now do it for brain-inspired computing?

X-ray spectromicroscopy investigation of soft and hard breakdown in RRAM devices.

Resistive random access memory (RRAM) is considered an attractive candidate for next generation memory devices due to its competitive scalability, low-power operation and high switching speed. The technology however, still faces several challenges that overall prohibit its industrial translation, such as low yields, large switching variability and ultimately hard breakdown due to long-term operation or high-voltage biasing. The latter issue is of particular interest, because it ultimately leads to device failure. In this work, we have investigated the physicochemical changes that occur within RRAM devices as a consequence of soft and hard breakdown by combining full-field transmission x-ray microscopy with soft x-ray spectroscopic analysis performed on lamella samples. The high lateral resolution of this technique (down to 25 nm) allows the investigation of localized nanometric areas underneath permanent damage of the metal top electrode. Results show that devices after hard breakdown present discontinuity in the active layer, Pt inclusions and the formation of crystalline phases such as rutile, which indicates that the temperature increased locally up to 1000 K.

The interaction of gold and silver nanoparticles with a range of anionic and cationic dyes.

We describe the synthesis of charge-stabilised gold and silver nanoparticles by a modified Turkevich method and their interaction with a selection of cationic and anionic dyes. It was found that gold nanoparticles interact strongly with cationic dyes and in some cases enhanced absorption was observed by UV-visible spectroscopy. It is also shown that addition of cationic dyes to gold nanoparticles triggers aggregation of the nanoparticles into large, micrometre-scale clusters. Simultaneous fragmentation and agglomeration of the gold nanoparticles was observed at high concentrations of cationic dye in the solution. These effects were not observed when gold nanoparticles were mixed with anionic dyes, nor for silver nanoparticles with either cationic or anionic dyes.

Towards population inversion of electrically pumped Er ions sensitized by Si nanoclusters.

This study reports the estimation of the inverted Er fraction in a system of Er doped silicon oxide sensitized by Si nanoclusters, made by magnetron sputtering. Electroluminescence was obtained from the sensitized erbium, with a power efficiency of 10(-2)%. By estimating the density of Er ions that are in the first excited state, we find that up to 20% of the total Er concentration is inverted in the best device, which is one order of magnitude higher than that achieved by optical pumping of similar materials.

Committee machines-a universal method to deal with non-idealities in memristor-based neural networks.

Artificial neural networks are notoriously power- and time-consuming when implemented on conventional von Neumann computing systems. Consequently, recent years have seen an emergence of research in machine learning hardware that strives to bring memory and computing closer together. A popular approach is to realise artificial neural networks in hardware by implementing their synaptic weights using memristive devices. However, various device- and system-level non-idealities usually prevent these physical implementations from achieving high inference accuracy. We suggest applying a well-known concept in computer science-committee machines-in the context of memristor-based neural networks. Using simulations and experimental data from three different types of memristive devices, we show that committee machines employing ensemble averaging can successfully increase inference accuracy in physically implemented neural networks that suffer from faulty devices, device-to-device variability, random telegraph noise and line resistance. Importantly, we demonstrate that the accuracy can be improved even without increasing the total number of memristors.

Isolation of Aleutian mink disease virus by affinity chromatography.

Affinity chromatography was used to isolate the Aleutian disease virus of mink. Dissociation of the immunoadsorbent-virus complex with 0.75 molar sodium chloride and then with a glycine-hydrochloride gradient released infective particles resembling picornaviruses. The elution profile suggests that two different types of virus-antibody complexes are formed, one dissociated by sodium chloride and another that requires glycine-hydrochloride in addition to sodium chloride for release of virus.

Intrinsic Resistance Switching in Amorphous Silicon Suboxides: The Role of Columnar Microstructure.

We studied intrinsic resistance switching behaviour in sputter-deposited amorphous silicon suboxide (a-SiO x ) films with varying degrees of roughness at the oxide-electrode interface. By combining electrical probing measurements, atomic force microscopy (AFM), and scanning transmission electron microscopy (STEM), we observe that devices with rougher oxide-electrode interfaces exhibit lower electroforming voltages and more reliable switching behaviour. We show that rougher interfaces are consistent with enhanced columnar microstructure in the oxide layer. Our results suggest that columnar microstructure in the oxide will be a key factor to consider for the optimization of future SiOx-based resistance random access memory.

Lymphoproliferative diseases of fowl: JM-V leukemic lymphoblasts in cell culture.

JM-V leukemic lymphoblasts were established in cell culture. The cultured cells (JM-VLC cells) were transplantable in young chicks and produced a disease indistinguishable from JM-V lymphoblastic leukemia as initiated by whole-blood inoculation. JM-VLC cells maintained a normal female karyotype through 13 passages in Rhode Island Red cockerels. With the use of JM-V antisera and antisera from birds with naturally occurring Marek's disease (MD), specific antigens were detected on the surfaces of living cells. Intracellular antigens were detected with anti-MD virus sera after cultivation for at least 1 day at 37 degrees C. In spite of the expression of MD antigens, the presence of herpesvirus particles associated with the cultured cells, and the occurrence of foci of multinucleated cells in kidney cultures from chicks inoculated with cellfree preparations of JM-VLC cells, the pathologic potential of the cultured cells was that of JM-V leukemia.

Lymphoproliferative diseases of fowl: chromosome breaks caused in lymphocytes by JM-V herpesvirus.

Chromosome preparations were made of bone marrow cells and peripheral lymphocytes isolated from chicks that developed leukemia following infection with JM-V herpes-virus. Karyotypic analysis revealed a high frequency of chromosome breaks and aneuploidy, as well as some pulverization of chromosomes. The number of chromosome breaks began to increase at 2-3 days post infection, and by 5 days post infection it reached 12.7% of bone marrow cells and 17.2% of peripheral lymphocytes. Similarly, the number of aneuploid metaphase figures increased rapidly and reached 12% of bone marrow cells and 19% of peripheral lymphocytes at 5 days post infection. Some specificity was observed in the chromosomes that were affected.

Lymphoproliferative diseases of fowl: characterization of transplantable G-B1 Marek's disease tumor cells in culture.

A transplantable Marek's disease tumor, derived and maintained through in vivo passage in syngeneic G-B1 chickens, was established in cell culture and characterized. The cells (GBT cells) grew in suspension and appeared morphologically similar to other Marek's lymphoblastoid lines except for prominent nucleoli in GBT cells. The number of chromosomes increased during in vitro cultivation from near triploid to subtetraploid. Chromosome no. 3 was abnormally short. The GBT cells retained surface B1 histocompatibility antigen of the original tumor as well as Marek's disease tumor-specific surface antigen (MATSA); however, infectious Marek's disease virus was not detected by immunofluorescence, electron microscopy, inoculation into susceptible birds, or assay on susceptible cells in culture. Inoculation of as few as 100 cultured cells into syngeneic G-B1 chickens results in tumor formation at the inoculation site, with eventual death of the recipients.

Delayed wound healing in mice associated with viral alteration of macrophages.

Mice acutely and chronically infected with Sendai virus had impaired healing of incisional wounds. This impairment could be overcome by instilling the macrophage-stimulators zymosan and glucan into full-thickness incised wounds, whereas levamisole, an immunomodulator, had no effect on tensile strength of these wounds (breaking strengths). Another commonly occurring murine virus, murine hepatitis virus, also reduced wound tensile strength in infected mice. However, the murine strain of herpes simplex virus, type 1, which caused greater morbidity, did not reduce tensile strength. Although measurements of wound length and wound breaking strength clearly showed the adverse effect of viral infection, histologic evaluation of wounds did not consistently reveal a similar change in composition of wound "cell aggregation centers" for mice infected with the Sendai and murine hepatitis viruses.

Comparison of early splenic changes associated with replication of viruses in murine monocytic macrophages.

An effect of replication of certain viruses in murine monocytic macrophages was manifested by depletion of cells through degenerative and necrotizing changes in thymus-dependent areas of lymphoid structures. In mice infected with murine hepatitis virus (MHV-3) or lactate dehydrogenase virus, these changes were transient in mice killed on postinoculation day (PID) 2. To study these morphologic changes due to viral replication, adult Swiss specific-pathogen-free homozygous nude mice (nu/nu) and their heterozygous haired littermates (nu/+) were inoculated with 10(5) LD50 of MHV-3, euthanatized, and necropsied on PID 1, 2, 4, 6, 8, and 10 along with noninoculated controls. The nu/+ and nu/nu mice killed on PID 2 had lymphocytic karyorrhexis and depletion of cells in the thymus-dependent area. In the heterozygote, these characteristic lesions were transient; whereas in the homozygote, lesions persisted and were present in survivors euthanatized and necropsied on PID 16. Although the intensity of lesions due to MHV-3 varied between nu/+ and nu/nu mice, virus titers determined on liver homogenates were similar for the homozygote and heterozygote during acute disease. Nude and nonnude mice given lactate dehydrogenase virus and killed on PID 2 had a transient depletion of lymphocytes; whereas mice given lymphocytic choriomeningitis virus and killed on PID 4 had a similar lesion. Lesions neither occurred when mice were treated with silica before inoculation, indicating that functional monocytic macrophages were required, nor occurred when another virus, herpes simplex virus type 1, was given.

Modulation of early cellular events in wound healing in mice.

Wound healing experiments were conducted in random-bred Swiss mice to determine the effect on antimicrobial surfactants and macrophage stimulators on wound measurements, histologic repair, and wound breaking strengths. In selection of mice, Sendai virus antibody-positive mice healed more slowly than did mice with no detectable titer to Sendai virus. Studies were conducted with Sendai virus-free mice that had C31G (an antimicrobial surfactant), alkyl amine oxide, zymosan, glucan, or phosphate-buffered isotonic saline solution instilled into full-thickness incised wounds. The early events in the repair process indicated a greater degree of inflammatory response comprised mainly of polymorphonuclear leukocytes with subsequent large numbers of monocytes in C31G and alkyl amine oxide-treated wounds. Although zymosan did not induce as large a number of monocytes, the degree of fibroplasia was as great as in wounds in which numbers were higher. The effect of zymosan could be blocked by the addition of N-alpha-p-tosyl-L-lysine chloromethyl ketone to wounds. Wound breaking strength 3 days after surgery was greatest for glucan-treated mice (134 +/- 37 g) whereas that in C31G-treated mice (77 +/- 31 g) was less than that of the controls (92 +/- 37 g). By day 7, there was no significant difference in breaking strength between control and glucan-treated wounds; however, C31G-treated wounds remained substantially weaker than control wounds.

Histamine-induced suppressor macrophage inhibits fibroblast growth and wound healing.

To determine whether there is histamine-induced suppressor activity in macrophage-related functions other than in immunity, extracts and media from a macrophage cell line, RAW 264, were tested for suppressor effect on fibroplasia. The procedure consisted of priming confluent RAW 264 cells in culture with media or cellular extracts of washed mastocytes (P-815). The inoculum was removed from the RAW 264 cells by rinsing with fresh medium 24 hours later, and then with medium replacement and 3 more days of culture. The culture media or extracts of washed RAW 264 cells were tested for suppressor activity. The primed RAW 264 cells were lysed by 4 freeze-thaw cycles and cleared by centrifugation, and the resulting supernatant was tested on fibroblast (3T3) cell growth and wound healing in mice and for suppressor activity on T cells. Replication of 3T3 cells, as quantitated by uptake of [3H]thymidine, was reduced 75% when "suppressor" material from RAW 264 cells was added to 3T3 cultures and not when media or extracts of unprimed RAW 264 cells were added. Tensiometric measurements of wound breaking strength (full-thickness incised wounds) were reduced 31% by day 4 and 47% by postsurgical day 7 when "suppressor" RAW 264 extracts were instilled into wounds. Leukocyte cultures stimulated with phytohemagglutinin had a reduced uptake of [3H]thymidine (suppressed 90% to 95%) when exposed to primed RAW 264 extracts, whereas kidney cell culture lines were unaffected. The data obtained indicated that mastocyte (histamine)-induced suppressor factors are present for fibroblast activity as well as T-cell function.

Characterization of deoxyribonucleic acid from cells infected with Aleutian disease virus.

Viral DNA was extracted from Crandell feline kidney (CRFK) cells infected with Aleutian disease virus (ADV) and labeled with [ 3H ]thymidine. The sedimentation coefficient in alkaline sucrose gradients was 16S corresponding to a molecular weight of 1.5 X 10(6). The buoyant densities of DNA from infected and control cells were determined by isopyknic sedimentation in CsCl and NaI gradients. Two additional peaks of [ 3H ]DNA were found in infected cells, but not in control cell extracts. Fractionation of this DNA on hydroxylapatite indicated that the new peaks represented a single-stranded component, density 1.728 g/cm3, and a double-stranded component, presumed to be a viral replicative intermediate, density 1.718 g/cm3. The target antigen formation in CRFK cells was measured by gamma-irradiation of ADV and assayed for focus formation. The calculated size of ADV based on these measurements was 1.1 X 10(6). The H-1 parvovirus also was shown to have a size of 1.5 X 10(6) daltons for both antigen and plaque formation. The data indicated similarities existed between ADV and other autonomously replicating parvoviruses in most properties, except that less-than-unit length genome of ADV may be transcribed.

Protides of the Mustelidae: immunoresponse of mustelids to Aleutian mink disease virus.

Members of North American Mustelidae were tested for their response to inoculation with 10(6) infective doses of Aleutian disease virus. In subfamily Mustelinae, 3 species in the genus Mustela (M vision, M erminea, and M putorius) and 2 species in genus Martes (Ma pennanti and Ma americana) responded immunologically with some features resembling Aleutian disease in mink. In subfamily Mephitinae, only Mephitis mephitis responded, and others of the subfamily did not, nor did members of subfamilies Melinae and Lutrinae. The responses observed ranged from development of detectable antibody levels determined by counterimmunoelectrophoresis to histopathologic changes typical of Aleutian disease.

Effect of C31G, an antimicrobial surfactant, on healing of incised guinea pig wounds.

An anti-infective surfactant composition (C31G) promoted healing of infected and noninfected wounds in guinea pigs. In this animal model, histologic features of wounds treated topically with C31G revealed an increased rate of wound closure associated with decreased inflammatory response and increased C31G fibroblast infiltration and epithelialization. The effect of C31G on fibrin formation, the initial event of wound healing, was compared with effects of anionic and cationic surfactants that delay healing. The surfactants had different effects on clotting time, platelet activation, and cross-linkage of the stabilized clot. Seemingly, C31G increased the protein cross-linking of fibrin in clots containing fibronectin.

Comparison of antipseudomonad activity of chlorine dioxide/chlorous acid-containing gel with commercially available antiseptics.

A chlorine dioxide-containing gel was compared with 3 commercially available antimicrobials and 1 antibiotic intended for topical use. This gel was tested at 0.5 X and 4 X and was found to be more effective as a 4 X gel in treating Pseudomonas aeruginosa-infected excised wounds on mice. To further compare this gel with other antiseptics, a murine bioassay was developed. This wound model consisted of an excised cutaneous wound on the dorsum of mice which were irradiated (800 rad) and inoculated with P aeruginosa at 10-fold dilutions, from 10(-2) to 10(-10). The wounds were observed for latency of infection or mice survival time as a function of concentration of viable organisms remaining after treatment. The advantage of this model was demonstrated where a standard curve based on latency did not consume as many test subjects and yet provided an estimate of viable organisms in each wound. In this model, the chlorine dioxide-containing gel was more active than were preparations of providone-iodine, chlorhexidene, or silver sulfadiazine and was similar to polymyxin-bacitracin-neomycin ointment as a topical antiseptic. The effectiveness of the tested gel was reduced if delays in treatment were longer than 1 hour.

Controlled wound repair in guinea pigs, using antimicrobials that alter fibroplasia.

The repair processes of incised wounds depend, in part, on fibroplasia induced by soluble mediators from monocytic macrophages. Two topical antimicrobials were evaluated, each of which effectively controlled wound sepsis and yet each had widely different effects on fibroplasia and wound strength. Paired-incision dermal wounds on the flanks of guinea pigs were treated with a substance containing reactive chlorine (Alcide) or with a compound that is a mixture of two surfactants. One side of each guinea pig was treated with one of the antimicrobials (treated wounds); the opposite side was treated with isotonic saline solution (control wounds). At 7, 10, and 16 days after surgery, tensiometric measurements of C31G (a surfactant)-treated wounds were 99%, 139%, and 195% of control wound values, respectively. Alcide-treated wounds were 76%, 58%, and 88% of control wounds, respectively. Wounds treated with chlorhexidine had reduced strength at 7 days (73%) and at 10 days (78%), but by 14 days, they were similar to control wounds (94%). The main difference between the wounds was the amount of collagen formation. Alcide-treated wounds incorporated less than 50% of the amount of 14C-proline than did the wounds treated with C31G. However, Alcide-treated wounds epithelialized as rapidly as did control wounds, and had minimal scar formation. Microscopic evaluations indicated greatly reduced inflammatory infiltrates in Alcide-treated wounds, indicating that reduced wound strength may be associated with lack of fibroblast-stimulating activity by monocytes.

Mice and rabbit models for oral and percutaneous absorption and disposition of amphoteric surfactant C31G.

A topical antimicrobial agent (C31G), composed of amphoteric surfactants (alkyl betaines and alkyl amine oxides), had a significant (P less than 0.001) and concentration dependent rate of percutaneous absorption in mice and rabbits. Percutaneous absorption studies (3 concentrations for 3 exposure periods) in mice indicated that dermal interaction resulted in nonlinear changes for rate of penetration and transport as a function of exposure and concentration. Kinetic studies of absorption and disposition after oral dosing in mice with [3H]C31G were used to determine the body burden (0.21 g/kg of body weight) at a no effect level. Dermal penetration (P) and transport (T) resulting from percutaneous exposure at variations of time and dosage concentrations were shown to fit the equation T = a + k ln t. The regression coefficient k represents the characteristic change of T with exposure time (t). For low concentrations k equaled 1.89 and at the high concentration k equaled 2.68. This increase of T indicated the interaction of the agent with the dermal barrier at high concentrations. Dermal transport in the rabbit was less than one-fourth of that in the mouse. Excretion of C31G after oral or dermal dosing was predominantly renal at higher dosage levels, whereas fecal excretion dominated at the lowest dosage levels. The half-life of [3H]C31G in the mouse was 68 hours and 86 hours in the rabbit. Data and statistical methods allowed prediction of the effects of daily exposures