Nature Studies Show 40% of fMRI Scans Misread Human Brain Activity

Sabine Hossenfelder

Perverse academic and publishing incentives drive researchers to ignore flawed measurement methods, leaving decades of foundational brain-mapping studies vulnerable to a replication collapse.

AI-assisted summary of Neuroscience is in big trouble — original source. Check the original for full context.

Source published: 2026-08-15

Key ideas

A study in Nature Neuroscience revealed that roughly 40% of functional MRI signals do not reflect neural activity as long assumed, as regions like the default mode network can consume oxygen directly without triggering an increase in blood flow. (0:00)
A meta-analysis of lesion network mapping covering over 100 neurological and psychiatric conditions found that supposedly disease-specific brain circuits appear virtually identical across distinct disorders, indicating non-specific artifacts. (1:46)
Cross-manufacturer imaging trials showed that individual functional brain network maps cannot be reliably replicated when switching between MRI scanners from different vendors like Siemens and Philips. (2:15)
Neuroscience is repeating the replication failures of psychology, driven by perverse incentives where researchers advance careers on flawed studies and publishers profit from sheer publication volume. (3:22)

Detailed notes

## Background Published on YouTube by Sabine Hossenfelder on August 12, 2026, this video examines a series of critical studies revealing fundamental measurement errors across brain research. According to the page description, neuroscientists have long relied on the assumption that increased brain activity drives increased blood flow in fMRI scans, but new research indicates this foundational premise is incorrect for roughly 40% of signals.

## TL;DR Sabine Hossenfelder details a deepening credibility crisis in neuroscience, headlined by a study showing that approximately 40% of fMRI signals mischaracterize neural activity because regions like the default mode network can consume oxygen without increasing blood flow. Further research reveals that lesion network mapping frequently generates non-specific circuits shared across unrelated disorders, while brain scans of the same subject vary drastically depending on the scanner's manufacturer. Hossenfelder argues that much like psychology's replication crisis, these systemic flaws persist because perverse career and publishing incentives reward quantity over methodological rigor.

## Atmosphere & style The presentation is trademark Sabine Hossenfelder: brisk, deadpan, and punctuated with biting, cynical humor. Moving from alarming methodological exposés to dry analogies about car tires and bad haircuts, she treats academic complacency with clinical exasperation. The tone feels like an irreverent yet deeply informed debriefing that punctures the prestige of modern neuroimaging with sharp, unflinching critique.

## Bulletin - [00:00](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=0s) A study in Nature Neuroscience revealed that roughly 40% of functional MRI signals do not reflect neural activity as long assumed, as regions like the default mode network can consume oxygen directly without triggering an increase in blood flow. - [01:46](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=106s) A meta-analysis of lesion network mapping covering over 100 neurological and psychiatric conditions found that supposedly disease-specific brain circuits appear virtually identical across distinct disorders, indicating non-specific artifacts. - [02:15](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=135s) Cross-manufacturer imaging trials showed that individual functional brain network maps cannot be reliably replicated when switching between MRI scanners from different vendors like Siemens and Philips. - [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) Neuroscience is repeating the replication failures of psychology, driven by perverse incentives where researchers advance careers on flawed studies and publishers profit from sheer publication volume.

## Key takeaways - [00:00](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=0s) Recent research published in Nature Neuroscience demonstrates that functional magnetic resonance imaging (fMRI) misinterprets about 40% of brain signals. - [00:35](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=35s) The traditional fMRI assumption that active neurons necessarily trigger increased regional blood flow fails in regions like the default mode network, which can extract more oxygen without increasing blood flow. - [01:20](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=80s) Statistical vulnerabilities in fMRI analysis have been recognized since at least 2009, when researchers famously detected apparent brain activity in a dead Atlantic salmon due to uncorrected false positives. - [01:46](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=106s) Lesion network mapping, a method intended to link brain damage patterns to specific symptoms, generates non-specific circuits that look almost identical across more than 100 different psychiatric and neurological conditions. - [02:15](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=135s) Individual functional brain maps vary significantly depending on whether a patient is scanned using a Siemens or a Philips MRI machine, preventing basic cross-hardware replication. - [02:40](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=160s) Exposing flawed neuroscience methodology has accumulated so much critical research that it has effectively become its own distinct academic subfield. - [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) Despite widespread awareness of these methodological problems for over 15 years, researchers routinely ignore them because academic career progression rewards published output over procedural rigor. - [04:10](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=250s) The crisis in neuroscience closely mirrors the psychology replication crisis, where underpowered studies and flawed statistical techniques invalidated decades of foundational literature. - [04:40](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=280s) Commercial incentives compound the problem because academic publishers generate revenue based on the volume of papers published rather than the reliability of findings. - [05:15](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=315s) Because the flawed methods underpin the primary goal of modern neuroscience—mapping distinct brain regions to specific functions—the reliability of the broader discipline is undermined. - [05:40](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=340s) Severe reproducibility gaps and scanner inconsistencies suggest that consumer-ready brain-computer interfaces remain significantly farther away than commonly projected.

## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=0s) The fMRI Crisis: 40% of Brain Signals Are Misinterpreted

Sabine opens with a blunt assessment: "Neuroscience is not doing well. Not at all." Over the past several months, multiple independent research groups have published studies demonstrating that a vast swath of research into the human brain and human behavior is simply wrong. The most devastating of these recent papers comes from a research group in Germany and was published in *Nature Neuroscience*. The team investigated brain studies that rely on functional magnetic resonance imaging (fMRI) and discovered that roughly 40% of these fMRI signals do not actually measure what neuroscientists have long assumed they measure.

Traditionally, fMRI scans work by detecting changes in blood flow. Scientists operated on the assumption that increased blood flow to a brain region means neurons there are consuming more oxygen and firing more rapidly—and therefore, a stronger fMRI signal directly translates to greater brain activity. However, by studying over 40 healthy volunteers using a more advanced metric for blood oxygen consumption, the German researchers found that this core assumption frequently breaks down. Specifically, in regions like the brain's "default mode network"—which is active during daydreaming or memory recall—neurons can ramp up their activity simply by extracting more oxygen directly from the existing blood supply without triggering any increase in blood flow. Because standard fMRI misses this or misinterprets the dynamic, the conclusions of thousands of fMRI studies conducted over the past several decades may be misleading.

```mermaid flowchart TD accTitle: Traditional fMRI Assumption vs. New Nature Neuroscience Findings accDescr: Compares the classic assumption that neural activity always increases blood flow with the new finding that regions like the default mode network can extract more oxygen without changing blood flow.

subgraph Classic["Traditional fMRI Assumption"] A1["Neurons Fire More"] --> A2["More Oxygen Used & Blood Flow Increases"] A2 --> A3["Larger fMRI Signal Detected"] end

subgraph Actual["New Finding (e.g., Default Mode Network)"] B1["Neurons Become More Active"] --> B2["Pull More Oxygen Directly From Blood"] B2 --> B3["No Increase in Blood Flow (~40% of signals misled)"] end ```

Mentioning the default mode network prompts Sabine to recall a classic cautionary tale from 2009: researchers placed a dead Atlantic salmon inside an fMRI scanner, showed it photographs of humans in social situations, and—thanks to uncorrected statistical errors—detected apparent "brain activity" in the deceased fish.

[01:46](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=106s) Flawed Lesion Network Mapping and Scanner Discrepancies

"But wait, it gets worse," Sabine continues. Just a month after the fMRI study, a second critical paper appeared in *Nature Neuroscience*, this time scrutinizing "the methodological foundations of lesion network mapping." Lesion network mapping is a technique used to study patients with brain damage who exhibit identical symptoms. Researchers look for commonalities across the damaged areas, and if those separate lesions connect to the same broader brain circuit, scientists conclude that this wider network is responsible for the symptom. However, when the authors of the new paper reviewed studies covering more than 100 neurological and psychiatric conditions, they discovered that many supposedly disease-specific brain circuits looked remarkably similar across completely different disorders. As the authors concluded, "the current results suggest that a substantial proportion of the presented lesion network mapping are non-specific and may not accurately reflect genuine biological brain networks." Sabine dryly compares the methodology to "finding that all broken cars have tires and concluding that tires cause engine failure, which by all chances would still get citations."

Compounding the crisis, another research group reported in April that scanning people's brains with MRI machines from different manufacturers yields dramatically different results. The discrepancies were so severe that researchers could not even replicate the functional brain network maps of individual subjects across machines—meaning your brain map "depends on whether you ask Siemens or Philips." Sabine notes that these aren't even the only recent papers flagging systematic flaws in neuroscience; so many critiques have emerged lately that exposing bad neuroscience methods has practically become a subfield in its own right.

| Study / Timeframe | Target Method / Equipment | Core Finding | |:--- |:--- |:--- | | **German Group (*Nature Neuroscience*)** | Functional MRI (fMRI) blood-flow signals | ~40% of signals don't measure what was assumed; neurons (e.g., in the default mode network) can extract more oxygen without increasing blood flow. | | **Follow-up Study (*Nature Neuroscience*, 1 month later)** | Lesion network mapping (across >100 conditions) | Supposedly disease-specific circuits look nearly identical across very different neurological and psychiatric disorders, indicating non-specific artifacts. | | **April Study** | MRI scanners from different manufacturers | Individual functional brain network maps could not be replicated when switching between manufacturers like Siemens and Philips. | | **2009 Salmon Study** | fMRI statistical analysis | A deceased Atlantic salmon shown photos of humans registered false-positive "brain activity" due to statistical errors. |

[03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) Perverse Incentives and the Psychology Replication Parallel

Faced with this avalanche of flawed research, Sabine asks, "So what the hell is going on?" While one could optimistically view these exposés as the healthy process of science correcting itself, she argues we should be asking why these errors happen in the first place—and why they keep happening. Pointing back to the 2009 dead salmon study, she emphasizes that these methodological issues have been known for more than 15 years. While it is great that critical researchers are publishing corrective papers, the sheer volume of flawed literature proves that most studies in these areas blatantly ignore well-known problems with data collection, measurement interpretation, and statistical analysis—"and they will continue to ignore them for as long as they can get away with it."

After a brief pause to thank her Patreon and YouTube supporters (especially those in Tier 4 and higher), joking that "unlike some brain signals, your support is very easy to interpret," Sabine draws a direct parallel between neuroscience and psychology. For decades, psychologists squeezed bogus results out of underpowered studies using statistical techniques that were widely known to be flawed. Today, supposedly foundational psychology studies fail replication constantly, to the point where "basically the entire field of psychology has to start over again."

The root cause in both fields is a system of perverse incentives: - **Scientists** have overwhelming career incentives to ignore methodological flaws so long as they can get their papers published. - **Scientific publishers** have strong financial incentives to keep publishing those papers because publication volume is how they make money.

"The tragedy is that this problem isn't remotely new," Sabine remarks. "It's widely known and yet nothing changes, which come to think of it is also a very good summary of my relationship with my hair." She then tries to offer a diplomatic caveat—"Now look, I don't want to say all of neuroscience is bullshit. Certainly not"—by noting that these critical papers only target specific subfields focused on identifying which brain regions or connections perform specific functions. She immediately deadpans: "Which, yeah, it's basically what neuroscience is about, isn't it?" Ultimately, she suspects these foundational reliability problems mean we are still a long way off from working brain-computer interfaces, "because it would be nice if Siemens and Philips can agree whether you actually wanted to post that image on Twitter."

[06:11](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=371s) Sponsor Segment: Novium Hoverpen

Transitioning to the video's sponsor, Sabine highlights that "good design is art combined with science" and introduces the Hoverpen by Novium. The pen stands vertically and floats in its base using strong permanent magnets, allowing it to spin freely. Because it is precisely balanced, users don't have to fumble to set it into the base, and it won't fall out on its own.

- **Hoverpen Interstellar:** Balanced at the exact angle of Earth's axial tilt, with a premium version available that features a genuine piece of outer-space meteorite embedded in the pen. - **Hoverpen Future:** Features a slightly different design equipped with interchangeable fountain and rollerball tips. - **Usability & Offer:** Both versions are refillable and flow smoothly across the page. Viewers can get 15% off for the first 72 hours (and 10% off afterward) by using the QR code on screen or the promo code `SABINE`.

## Notable quotes > [00:00](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=0s) Neuroscience is not doing well. Not at all. — Sabine Hossenfelder

> [01:46](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=106s) the current results suggest that a substantial proportion of the presented lesion network mapping are non-specific and may not accurately reflect genuine biological brain networks. — Sabine Hossenfelder

> [01:46](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=106s) finding that all broken cars have tires and concluding that tires cause engine failure, which by all chances would still get citations. — Sabine Hossenfelder

> [01:46](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=106s) depends on whether you ask Siemens or Philips. — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) So what the hell is going on? — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) and they will continue to ignore them for as long as they can get away with it. — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) unlike some brain signals, your support is very easy to interpret — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) basically the entire field of psychology has to start over again. — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) The tragedy is that this problem isn't remotely new. It's widely known and yet nothing changes, which come to think of it is also a very good summary of my relationship with my hair. — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) Now look, I don't want to say all of neuroscience is bullshit. Certainly not — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) Which, yeah, it's basically what neuroscience is about, isn't it? — Sabine Hossenfelder

> [03:22](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=202s) because it would be nice if Siemens and Philips can agree whether you actually wanted to post that image on Twitter. — Sabine Hossenfelder

> [06:11](https://www.youtube.com/watch?v=jHGx7HBB-J4&t=371s) good design is art combined with science — Sabine Hossenfelder

## Mentioned Sabine Hossenfelder, Nature Neuroscience, Siemens, Philips, Patreon, YouTube, Twitter, Novium, Hoverpen, Hoverpen Interstellar, Hoverpen Future

## Video description For 72 hours, enjoy 15% OFF on all Hoverpens with code SABINE, or click on the link https://noviumdesign.shop/sabine - Free shipping to most countries. Also on Amazon: https://noviumdesign.shop/jTAp1f

Neuroscientists have long assumed that when a brain region becomes more active, more blood flows to it. That assumption is the entire basis of functional magnetic resonance imaging. A new paper now says it's wrong for about 40% of the signals, and it isn't the only recent study to find that a lot of what we thought we knew about the brain rests on bad measurements. Let's take a look.

#science #neuroscience #sciencenews #brain #health

👕T-shirts, mugs, posters and more: ➜ https://sabines-store.dashery.com/ 💌 Support me on Donorbox ➜ https://donorbox.org/swtg 👉 Transcript with links to references on Patreon ➜ https://www.patreon.com/Sabine 📝 Transcripts and written news on Substack ➜ https://sciencewtg.substack.com/ 📩 Free weekly science newsletter ➜ https://sabinehossenfelder.com/newsletter/ 👂 Audio only podcast ➜ https://open.spotify.com/show/0MkNfXlKnMPEUMEeKQYmYC 🔗 Join this channel to get access to perks ➜ https://www.youtube.com/channel/UC1yNl2E66ZzKApQdRuTQ4tw/join 📚 Buy my book ➜ https://amzn.to/3HSAWJW

[Source](https://www.youtube.com/watch?v=jHGx7HBB-J4)

中文简报

研究显示：fMRI扫描误读了40%的大脑活动信号数十年来的人脑定位研究建立在存在缺陷的生理学假设和不可复现的扫描仪硬件之上，但学术出版激励机制阻碍了该学科的自我革新。随着新研究揭露神经成像与脑图谱绘制技术的根本缺陷，神经科学正面临日益加剧的信任危机。《自然·神经科学》（Nature Neuroscience）发表的一项研究表明，约40%的fMRI信号脱离了血流动力学假设，尤其是在默认模式网络中，神经元在没有血流激增的情况下也会摄取氧气。更严重的是，病灶网络绘图在100多种不同的精神与神经疾病中得出了几乎完全相同的非特异性网络，且扫描结果在Siemens和Philips等不同硬件厂商的设备间无法复现。这些系统性缺陷与社会心理学的可复现性危机如出一辙，其根源在于学术出版激励机制长期忽视既有的方法学警告。《自然·神经科学》（Nature Neuroscience）的一项研究显示，约40%的fMRI信号并未反映局部血流激增，直接动摇了数十年来关于血流动力学变化与神经元放电存在关联的假设。 (0:00) 一项针对100多种精神和神经疾病的元分析显示，病灶网络绘图得出的主要是非特异性回路，在不同诊断之间看起来几乎完全相同。 (1:44) 在对相同受试者进行脑部扫描时，Siemens与Philips设备之间生成的功能网络图谱未能成功复现。 (1:44) 神经科学面临着系统性的可复现性危机，其根源在于学术出版的激励机制——尽管长期受到警告，该机制仍鼓励大量产出不可复现的脑定位与图谱论文。 (3:21)

Reading summary

Sabine Hossenfelder argues that core subfields of neuroscience face a profound reliability crisis, citing recent studies that expose widespread misinterpretations of fMRI data, unreplicable mapping techniques, and deep-seated publishing incentives that discourage methodological rigor. Flawed Assumptions in fMRI Signal Interpretation (0:00) A primary pillar of modern brain research—functional magnetic resonance imaging (fMRI)—rests on the assumption that heightened neural activity triggers increased blood flow. However, a study by German researchers published in Nature Neuroscience evaluated over 40 healthy volunteers using advanced blood oxygen metrics and found that roughly 40% of fMRI signals do not reflect what researchers assumed. In areas such as the default mode network, active during daydreaming or memory recall, neurons can draw additional oxygen directly from existing blood supplies without increasing blood flow, rendering standard fMRI signals potentially misleading across thousands of historical studies. Hossenfelder notes that analytical vulnerabilities in fMRI have been documented for years, citing a 2009 experiment in which uncorrected statistical errors produced apparent brain activity in a dead Atlantic salmon. Mapping Artifacts and Hardware Inconsistencies (1:46) Similar structural weaknesses undermine lesion network mapping, a technique used to identify shared circuits among patients with brain damage exhibiting identical symptoms. A subsequent review in Nature Neuroscience evaluated lesion network mapping studies spanning more than 100 neurological and psychiatric conditions, discovering that supposedly distinct, disease-specific brain networks appeared virtually identical across vastly different disorders. The authors concluded that many reported networks are non-specific artifacts rather than genuine biological structures. Adding to these discrepancies, an April study demonstrated that scanning the same individuals on MRI machines from different manufacturers, such as Siemens and Philips, generated functional brain network maps that could not be replicated across devices. Perverse Incentives and Replication Failures (3:22) Hossenfelder draws a direct parallel between the current state of neuroscience and the ongoing replication crisis in psychology, where underpowered studies and flawed statistical methods were tolerated for decades. She argues that these issues persist because the academic ecosystem provides perverse incentives: individual researchers face career pressure to overlook methodological flaws in order to secure publications, while scientific publishers profit from maximizing publication volume. While acknowledging that these critical findings focus specifically on the subfields dedicated to localizing brain functions and connections, Hossenfelder emphasizes that this work represents a central objective of neuroscience, suggesting that reliable applications such as brain-computer interfaces remain distant until basic measurement validity is resolved. Hossenfelder concludes that while the recent surge of critical papers demonstrates science's capacity for self-correction, the long-term neglect of known methodological defects indicates that neuroscience will continue producing flawed research until systemic publishing and career incentives change.

Source quotes

[0:00] Sabine Hossenfelder: Neuroscience is not doing well. Not at all. [1:46] Sabine Hossenfelder: the current results suggest that a substantial proportion of the presented lesion network mapping are non-specific and may not accurately reflect genuine biological brain networks. [1:46] Sabine Hossenfelder: finding that all broken cars have tires and concluding that tires cause engine failure, which by all chances would still get citations. [1:46] Sabine Hossenfelder: depends on whether you ask Siemens or Philips. [3:22] Sabine Hossenfelder: So what the hell is going on? [3:22] Sabine Hossenfelder: and they will continue to ignore them for as long as they can get away with it. [3:22] Sabine Hossenfelder: unlike some brain signals, your support is very easy to interpret [3:22] Sabine Hossenfelder: basically the entire field of psychology has to start over again. [3:22] Sabine Hossenfelder: The tragedy is that this problem isn't remotely new. It's widely known and yet nothing changes, which come to think of it is also a very good summary of my relationship with my hair. [3:22] Sabine Hossenfelder: Now look, I don't want to say all of neuroscience is bullshit. Certainly not [3:22] Sabine Hossenfelder: Which, yeah, it's basically what neuroscience is about, isn't it? [3:22] Sabine Hossenfelder: because it would be nice if Siemens and Philips can agree whether you actually wanted to post that image on Twitter. [6:11] Sabine Hossenfelder: good design is art combined with science

Original transcript

[0:00] Neuroscience is not doing well. Not at all. In the past months, we've seen multiple studies from [0:07] independent groups showing that a large number of studies about the human brain and human behavior [0:13] are just wrong. I have a brief summary and then I want to talk about what this means and how it [0:20] could happen. The most devastating recent paper comes from a group in Germany and was published in [0:26] Nature Neuroscience. They looked at brain studies which rely on a measurement technique called [0:31] functional magnetic resonance imaging, fMRI for short, and found that about 40% of these signals [0:38] simply don't measure what scientists thought they measure. These scans detect changes in blood flow. [0:46] Scientists assumed that more blood flow means that neurons use more oxygen and fire more. Hence, [0:54] a larger fMRI signal means more brain activity. But the new study shows that in regions like [1:00] the brain's "default mode network", which is active during daydreaming or memory recall, [1:06] neurons can become more active just by pulling more oxygen from the blood without increasing [1:13] blood flow. The researchers figured that out by a study on over 40 healthy volunteers, [1:19] which also used a more advanced measure for blood oxygen consumption. This means that the results of [1:25] thousands of studies done in the past decades might be misleading. My default mode network [1:32] brings up a story from 2009 where researchers placed a deceased Atlantic salmon in an fMRI [1:38] scanner, presented it with photos of humans in social situations, and detected apparent brain [1:44] activity due to statistical errors. But wait, it gets worse. Just a month later, a second paper [1:50] appeared in Nature Neuroscience. The authors looked at "the methodological foundations of [1:56] lesion network mapping". Lesion network mapping investigates patients with brain damage who have [2:02] the same symptoms. Scientists look for what the damaged places have in common and if those places [2:07] are connected to the same wider brain system, they claim that system is responsible for the symptom. [2:14] But the authors of the new paper reviewed studies for more than 100 neurological and psychiatric [2:21] conditions and found that many supposedly disease-specific circuits look highly similar [2:27] across very different disorders. They conclude the current results suggests that a substantial [2:33] proportion of the presented lesion network mapping are non-specific and may not accurately reflect [2:40] genuine biological brain networks. It's a bit like finding that all broken cars have tires [2:46] and concluding that tires cause engine failure, which by all chances would still get citations. [2:52] Then just in April, a group reported that if you scan people's brains with scanners from [2:58] different manufacturers, you get different results and not just a little bit different, [3:02] but they couldn't replicate the functional brain network map of individuals. Depends on whether [3:08] you ask Siemens or Phillips. These aren't even the only recent studies which flag systematic [3:15] problems with neuroscience studies. There are so many of these it's become somewhat of a sub field [3:21] in its own right. So what the hell is going on? On the one hand you can say it's a good development. [3:29] It's the process of science correcting itself. On the other hand you may ask and indeed I think [3:35] you should ask why the hell is this happening in the first place? And more important still, [3:41] why does it continue to happen? Because this isn't a new thing. As we saw, this has been [3:47] going on at least since 2009. And I know that's shocking, but that was more than 15 years ago. [3:54] It's important because you see the people who write these critical papers exist, which is great, [4:00] but it's also clear from the numbers that most of the studies that get published in the areas [4:05] ignore the known problems with the data collection, the measurement interpretation, [4:10] and the analysis, and they will continue to ignore them for as long as they can get away with it. [4:17] If you find my science news interesting, please consider supporting me on Patreon or right here on [4:23] YouTube. And if you already support me, especially those of you in tier 4 and higher, thank you so [4:30] much. This channel would not be possible without your help. And unlike some brain signals, your [4:36] support is very easy to interpret. The problem in neuroscience is extremely similar to what's [4:43] happened in psychology, where scientists have been squeezing fake results out of underpowered studies [4:50] with statistical techniques that were widely known to be wrong. This continued for decades and now [4:57] every time you turn around there is yet another supposedly foundational psychology study that [5:03] could not be replicated. Basically the entire field of psychology has to start over again. The [5:09] reason these things happen is that scientists have very strong incentives to ignore these problems [5:16] so long as they can get their papers published. and scientific publishers have strong incentives [5:22] to publish papers because that's what they make money from. The tragedy is that this problem isn't [5:29] remotely new. It's widely known and yet nothing changes which come to think of it is also a very [5:36] good summary of my relationship with my hair. Now look, I don't want to say all of neuroscience is [5:42] bullshit. Certainly not. The papers I just mentioned are in specific fields related in particular to [5:48] identifying certain regions or connections in the brain that do one or the other thing [5:53] which yeah it's basically what neuroscience is about isn't it? All of this makes me think [6:00] it'll probably be a while until we have our brain interfaces because it would be nice if Siemens [6:06] and Phillips can agree whether you actually wanted to post that image on Twitter. Good design is art [6:14] combined with science. And this pen from Novium is the perfect example of functionality with [6:21] eye-catching looks. The Hoverpens float in their base thanks to strong permanent [6:26] magnets and can freely spin. They're really well balanced. You don't need to fumble to put them in, [6:32] and they don't drop out either. This particular version is the Hoverpen Interstellar balanced at [6:38] the same angle as the tilt of Earth's axis. It has a premium version with a meteorite embedded. Yes, [6:45] a real meteorite. So, there's part of your pen that actually flew through outer space. [6:51] The Hoverpen doesn't just look good, it's also a pleasure to write with. It flows nicely across the [6:57] page and it's refillable, too. A slightly different design is the Hoverpen Future, [7:02] which has interchangeable tips, fountain, and roller ball. The Hoverpens are high quality [7:09] products that I've been using for years, and I'm happy to recommend them. And of course, [7:14] I have a special offer. You'll get 15% off in the next 72 hours and 10% after this if you use my [7:22] QR code or the code SABINE. Links in the info below. Thanks for watching. See you tomorrow.