Neuroscientist investigating the role of cerebral cortex and thalamus in our behavior

Oxford, UK
Randy Bruno retweeted
NEWS: Blood tests could identify early signs of cancer years before diagnosis A study of blood samples collected up to 9yrs before a cancer diagnosis found changes in the DNA that captured the body’s response to early traces of cancer development.  medsci.ox.ac.uk/news/blood-t…
6
12
816
Honored to meet the artist behind "Hearts & Minds", the great Patrick Hughes
3D "Reverse Perspective" of a floating stack of textbooks - amazing artwork "Hearts & Minds" by Patrick Hughes! His gift to Professor David Paterson in honor of David's 10 years leading our department @OxfordDPAG at @UniofOxford I love the crumpled edges of the jacket cover
1
1
7
653
3D "Reverse Perspective" of a floating stack of textbooks - amazing artwork "Hearts & Minds" by Patrick Hughes! His gift to Professor David Paterson in honor of David's 10 years leading our department @OxfordDPAG at @UniofOxford I love the crumpled edges of the jacket cover
1
1
854
Randy Bruno retweeted
When 90% of a time series is missing: reconstruct the dynamics, not the values Most methods for filling gaps in scientific time series try to predict the missing points from correlations in the observed data. Wu and coauthors take a very different route: recover the geometry of the dynamical system first. Using delay embedding, a single fully observed variable can reconstruct the system’s underlying attractor in state space. An incomplete variable gives a partial reconstruction of the same dynamics. The model then learns the mapping between those two reconstructed manifolds, using a Gaussian process to fill in the missing states. The clever part is that the machine learning is not asked to discover the dynamics from scratch. Dynamical-systems theory supplies the prior: different observations of the same system should encode equivalent underlying state-space structure. The Gaussian process only learns the mapping between those representations. That becomes especially powerful when data are extremely sparse. In a 120-dimensional Lorenz system, with only one variable fully observed and the other 119 missing 90% of their values, the recovered signals still reach an average correlation of 0.84 with the ground truth. A 120-dimensional Rössler system reaches 0.81. The same idea also transfers to real data. For 64-channel EEG, with one signal complete and the others 90% missing, reconstruction reaches correlations around 0.9. For AI for Science, the broader lesson is powerful: when your data come from a dynamical system, the missing values are constrained by the dynamics. Instead of asking ML to guess individual points, reconstruct the latent structure that makes those points possible. Wu et al., Nature Communications (2026), CC BY 4.0. doi.org/10.1038/s41467-026-7…
9
73
504
23,927
Randy Bruno retweeted
81
562
6,602
115,622
New translational neuroscience institute in Manhattan seeks lead
Chief Scientific Officer Search at the AOI The Ackman Oxman Institute (AOI) is a new translational neuroscience, rehabilitation, and brain health institute in New York City founded to accelerate the translation of promising science to treating and curing patients. We are seeking a Chief Scientific Officer (CSO) who shares our ambition and recognizes that transformative advances depend on creative, independent scientists pursuing fundamental questions relevant to brain injury, degeneration, recovery, and rehabilitation. The CSO will help shape AOI's scientific identity, organization, and culture. We welcome the candidate's ideas about how best to organize the science and create an environment in which outstanding investigators can do their best work. Mandate Lead AOI's discovery science enterprise. Working with the Chief Translational Officer (CTO) and institute leadership, the CSO will define the research agenda, recruit and develop exceptional scientists, and determine the mix of investigator-led laboratories, collaborative teams, shared resources, and external partnerships best suited to the mission. The portfolio will support two complementary kinds of work: Patient-Anchored Translational Programs Research developed in anticipation of, as well as in response to, clinical observations. For example, a mechanistic program on mitochondrial delivery and bioenergetic rescue could inform a parallel clinical study of mitochondrial transplantation. Programs pursuing a defined path to patients will have clear milestones and criteria for advancement, redirection, or stopping, appropriate to their stage of development. Mission-Relevant Discovery Science Protected support for fundamental research on brain injury, repair, plasticity, and aging, led by investigators with intellectual independence. Major continuation or stop decisions will use a five-to-seven-year horizon, with ongoing scientific exchange and peer review to maintain rigor. This work can generate new therapeutic ideas and give AOI the depth to recognize and pursue discoveries arising from patient studies. The CSO will shape the balance between these forms of research and how they interact, with resources and review practices suited to each. Scientists will have the freedom to pursue original ideas within AOI's mission and contribute to a collaborative scientific community. Scope and Responsibilities Scientific Portfolio Define the research agenda with the CTO across acquired brain injury (stroke, TBI), and neurodegenerative and neuropsychiatric diagnoses as well brain health in aging. Build on AOI's initial priorities in bioenergetics/mitochondria, neuroinflammation, and neuroplastogens (devices or biologics that increase neuroplasticity), helping refine their scope, sequence, and organization. Identify promising mechanistic programs early so AOI is prepared to act on clinical discoveries, and welcome new ideas through a transparent scientific review process. Independent Discovery Recruit exceptional investigators and give them the resources and latitude to pursue fundamental questions relevant to AOI's mission. Develop peer review that evaluates scientific quality, originality, and contribution over an appropriate time horizon, recognizes the value of well-reasoned negative results, and protects intellectual independence. Scientific Infrastructure Determine which laboratory and measurement capabilities will best support the science, and which to build internally, share, or access through partnerships. Guide scientific use of shared resources, including GLP vivarium and GMP capabilities, in coordination with the CTO and Chief Regulatory Officer. Scientific Recruitment and Community Shape a recruitment and support model that attracts outstanding scientists with diverse approaches to discovery and collaboration. Explore joint or affiliated academic appointments and shared recruitment with New York City and regional universities and medical centers, where these strengthen recruitment, mentorship, and scientific exchange. Appointment arrangements will be developed with partner institutions and individual recruits. External Scientific Network Build a network of well-funded collaborations with academic laboratories in New York City and the region whose fundamental work is relevant to AOI's mission, so that AOI's discovery capacity extends well beyond its own walls. The Pershing Square MIND Prize community is a natural springboard for this network, though collaboration is not limited to prize recipients. Work with the founders on how the MIND Prize and AOI's basic research program should relate to each other over time. Scientific Quality Lead a collegial, rigorous scientific review process. For translational and later-stage preclinical programs, agree on clear milestones and advancement or stop criteria with program leaders and the CTO. For fundamental discovery, combine ongoing scientific dialogue with substantive peer review and five-to-seven-year continuation decisions. Uphold research integrity and reproducibility across all programs. External Standing Represent AOI's science to the field, to funders (NIH, ARPA-H, DoD, philanthropy), to scientific advisors, and to academic and industry partners. Relationship to the Chief Translational Officer The CSO and CTO are peers who connect discovery with development and clinical investigation. The CSO leads fundamental and exploratory preclinical research; the CTO leads regulated preclinical development, manufacturing, engineering, and trial readiness, and clinical studies. They jointly define the evidence needed to advance a program into development and review the portfolio together. Scientific collaboration continues as programs advance and clinical observations raise new research questions. First 6 Months •Develop a scientific strategy with the CEO, CTO, and board, including the proposed research organization, the balance of discovery and translation, initial priorities and resources, and a first set of anticipatory mechanistic programs. Agree on review practices that protect independent discovery and support timely development decisions. •Begin recruiting priority scientific leaders and independent investigators, with the mix and sequence of appointments shaped by the scientific strategy. Develop the support and affiliation arrangements needed to attract exceptional candidates. •Establish an agreed plan for laboratory capabilities and launch initial research using in-house resources or external partnerships as appropriate. Work with the CTO and clinical leadership to align research with patient studies and their readiness requirements. •Co-define the criteria for entering development with the CTO and begin well-funded collaborations with regional academic laboratories. Set the subsequent launch timetable around scientific priorities, recruitment, and facility readiness. Who We Are Looking For Essential •MD/PhD or PhD with a first-rank research record in neuroscience, immunology/neuroinflammation, bioenergetics, or a convergent field. •Experience building and leading a research organization (institute, department, biotech R&D unit, or focused research organization). •A track record of attracting outstanding scientists. •Enthusiasm for a patient-centered research model and for leading a team whose work spans the continuum from discovery to therapy development. Valued •Experience supporting independent investigators and collaborative research in academia, research institutes, or industry. •Scientific credibility in at least one of AOI's initial priority areas. •Comfort spanning biologics and device/neurotechnology science. •A demonstrated ability to protect creative discovery while guiding translational programs toward clear milestones. Practical AOI welcomes candidates from academia and industry who are excited to build the institute over the coming decade and beyond. The position is full-time in New York City. Candidates should be eager to focus their research leadership on the brain, recovery, and rehabilitation; any continuing research commitments should support that mission. Compensation will be competitive and will include the opportunity to participate in the value created through company formation at AOI. Leadership Team Engagement Works closely with the CTO as described above and with clinical leadership to connect research and patient studies. Engages the legal/regulatory team early on regulatory strategy, quality systems, and evidence requirements. Collaborates with the Chief AI Officer on data-rich study design and computational modeling, the COO on laboratory development and recruitment operations, and the CFO on scientific resource planning. Interested candidates should send their CV and a brief cover note to: CSO@aoi.org
8
1,215
1/12 It has been quite a journey, and I am incredibly happy to finally share this work! One of the first projects we started when I opened my lab, with a brave team who took a chance on a very junior PI
10
22
59
6,433
Extracellular signatures of the different mediodorsal (MD) thalamus subnuclei
1
1
8
881
Randy Bruno retweeted
Morning coffee is linked to a longer life.
151
739
4,941
692,239
Randy Bruno retweeted
Excited to share our new Cell Reports paper on paratenial thalamus (PT), which identifies a molecular marker for PT and shows how PT interacts with the medial prefrontal cortex, nucleus accumbens, and basolateral amygdala, all done by my student Nigel Dao: sciencedirect.com/science/ar…
6
9
605
Randy Bruno retweeted
The Attention Deficit of AI Revealed Read more psychologytoday.com/us/blog/… Researchers discovered a weakness of LLMs using a common psychology test #innovation .@PsychToday .@QC_News .@TAMU #AI
2
2
331
Timeline Cleanse #25 (Kinder Scout, Peak District)
2
5
290
I’ve never seen anything more accurate in my life
webdivo nettinho
567
36,922
156,483
7,237,729
Randy Bruno retweeted
We're looking for a postdoc with NHP ephys experience to do simultaneous Neuropixel recordings across the corticothalamic motor circuit to better understand the neural control of reaching and grasping. We start looking at applications on August 31st. docs.google.com/document/d/1…
2
22
33
3,307
Randy Bruno retweeted
Absolutely insane stat: 15 U.S. colleges now cost more than $100,000 per year. That’s $400,000+ for a four-year degree. Students are not getting anywhere near $400,000 worth of value from these schools. Higher education is the most overpriced product in America.
833
1,028
6,217
1,287,213
Randy Bruno retweeted
We are pleased to announce our new interdisciplinary certificate program, Witchcraft with Data.
10
32
243
24,127
Randy Bruno retweeted
Reviewer 1: this could be two papers Reviewer 2: this should be no papers
26
281
3,366
99,578
Randy Bruno retweeted
A Stanford statistician built a dice-rolling machine to prove board game companies have been selling you loaded dice for a hundred years. The machine drops dice inside a glass chamber, records every landing, and runs the sequences through statistical tests. Every mass-produced die it tested was biased. Every Monopoly die. Every Yahtzee cube. Every D&D twenty-sider. Every casino die below one hundred dollars. The bias is so small the human eye cannot see it in one game. It is so consistent it drains a player over a night at a real table. His name is Persi Diaconis. He is 79. He ran away from home at fourteen to learn card magic from Dai Vernon. He came back twenty years later with a PhD in statistics from Harvard and became the person casinos, magicians, and mathematicians all quietly consult when they need to know how something actually shuffles, rolls, or falls. This eight-minute Numberphile clip is Diaconis explaining exactly what the machine found. The dots on a die are drilled holes. The paint filling them weighs less than the plastic that was removed. The six face is lighter than the one face. Every die you have ever owned has been tilted by a fraction of a gram since the day it left the factory. Casinos already knew. A single die at the Bellagio costs one hundred and fifty dollars, is discarded after eight hours of use, and is engineered so the paint in the pits weighs exactly what the plastic did. That is the whole difference between a game piece and a regulated instrument. Here is what Diaconis proved on the record. A cheap six-sided die favors the six face over thousands of rolls. A plastic twenty-sided die can favor certain numbers by a substantial margin depending on how the corners are cut. A thirty-sided die, unless the internal geometry is spherical, rolls some faces two to three times more often than others. One retired executive rolled a single die three and a half million times before it physically wore round. Euclid worked out part of this in 300 BC. Diaconis has spent forty years finishing the rest. "There is no such thing as a fair die. There are only dice fair enough for the purpose you are using them for." Fairness is not a property of the die. It is a property of the die plus the throw plus the surface. Change any one and the same die stops being fair. The video is free. Diaconis's papers are free on his Stanford page. Everyone who has ever argued over a roll has been trusting a piece of plastic they never checked. The math is free. The willingness to admit your die is loaded is a much rarer commodity than fairness.
39
173
3,003
296,699