Connect with us

News

Scientists use AI neural network to translate speech from brain activity

Published

7 years ago

on

Three recently published studies focused on using artificial intelligence (AI) neural networks to generate audio output from brain signals have shown promising results, namely by producing identifiable sounds up to 80% of the time. Participants in the studies first had their brain signals measured while they were either reading aloud or listening to specific words. All the data was then given to a neural network to “learn” how to interpret brain signals after which the final sounds were reconstructed for listeners to identify. These results represent hopeful prospects for the field of brain-computer interfaces (BCIs), where thought-based communication is quickly moving from the realm of science fiction to reality.

The idea of connecting human brains to computers is far from new. In fact, several relevant milestones have been made in recent years including enabling paralyzed individuals to operate tablet computers with their brain waves. Elon Musk has also famously brought attention to the field with Neuralink, his BCI company that essentially hopes to merge human consciousness with the power of the Internet. As brain-computer interface technology expands and develops new ways to foster communication between brains and machines, studies like these, originally highlighted by Science Magazine, will continue demonstrating the steady march of progress.

Functional areas of the human brain. | Credit: Blausen.com staff (2014) via CC BY 3.0.

In the first study conducted by researchers from Columbia University and Hofstra Northwell School of Medicine, both in New York, five epileptic participants had the brain signals from their auditory cortexes recorded as they listened to stories and numbers being read to them. The signal data was provided to a neural network for analysis which then reconstructed audio files that were accurately identified by participating listeners 75% of the time.

In the second study conducted by a team from the University of Bremen (Germany), Maastricht University (Netherlands), Northwestern University (Illinois), and Virginia Commonwealth University (Virginia), brain signal data was gathered from six patients’ speech planning and motor areas while undergoing tumor surgeries. Each patient read specific words aloud to target the data collected. After the brain data and audio data were given to their neural network for training, the program was given brain signals not included in the training set to recreate audio, the result producing words that were recognizable 40% of the time.

Finally, in a third study by a team at the University of California, San Francisco, three participants with epilepsy read text aloud while brain activity was captured from the speech and motor areas of their brains. The audio generated from their neural network’s analysis of the signal readings was presented to a group of 166 people who were asked to identify the sentences from a multiple choice test – some sentences were identified with 80% accuracy.

While the research presented in these studies shows serious progress towards connecting human brains to computers, there are still a few significant hurdles. For one, the way neuron signal patterns in the brain translate into sounds varies from person to person, so neural networks must be trained on each individual person. The best results require the best data possible, i.e., the most precise neuron signals possible, meaning this is something that can only be obtained by placing electrodes in the brain itself. The opportunities to collect data at this invasive level for research are limited, relying on voluntary participation and approval of experiments.

Advertisement

All three of the studies highlighted demonstrated an ability to reconstruct speech based on neural data in some significant capacity; however, also in all cases, the study participants were able to create audible sounds to use with the computer training set. In the case of patients unable to speak, the level of difficultly in interpreting the brain’s speech signals from other signals will be the biggest challenge. Also, the differences between brain signals during actual speech vs. thinking about speech will complicate matters further.

Related Topics:

Accidental computer geek, fascinated by most history and the multiplanetary future on its way. Quite keen on the democratization of space. | It's pronounced day-sha, but I answer to almost any variation thereof.

Advertisement
Comments

News

Tesla creates clever solution to simplify and improve its Service

Raj Jegannathan, a Vice President of IT/AI-Infra, Apps, Infosec, and Vehicle Service Operations, revealed that Tesla has started a small pilot program at a few service locations to combat this issue.

Published

7 hours ago

on

August 7, 2025

By

tesla service
Credit: Tesla

Tesla has created a clever solution to simplify and improve its Service. Tesla performs most of the services that are needed on its vehicles at its company-owned Service Centers.

However, service has been a weak point of the company, as some regions have fewer Service Centers than others. This can cause long wait times for Tesla owners in some parts of the country.

There are also instances where customers do not agree with what Tesla is saying about their vehicle. In fact, one instance that revealed this new change Tesla is making to its Service was precisely that.

One owner posted on X that his vehicle’s battery seal had failed after a recall was issued. Tesla insurance and Tesla Service both did not assist, and it took CEO Elon Musk stepping in to get the issue resolved:

Another owner suggested there should be a more streamlined communications process between the customer and the Service Center, a solution that has been missing.

Raj Jegannathan, a Vice President of IT/AI-Infra, Apps, Infosec, and Vehicle Service Operations, revealed that Tesla has started a small pilot program at a few service locations to combat this issue.

Elon Musk wants Tesla Service to fix two-thirds of cars in the same day

Jegannathan said that Tesla has started to share local and regional leader contact information so customers have the ability to reach out when they have complaints or disagree with warranty claims, changes in estimates, or initial diagnostics.

Advertisement

It is available in a handful of locations already, and Jegannathan said that once abuse guardrails are built, this will expand to all locations:

This would be a major improvement in the Service portion of Tesla’s business. There are common disagreements between Service and customers, specifically when Service’s suggestions don’t align with the customer’s beliefs.

When it comes to things like a warranty claim, these issues are not really up for interpretation. Instead, the repairs should be made. If there is a misunderstanding on Service’s side, a simple message from the customer could have resolved the issue. That’s basically what happened here.

Advertisement
Continue Reading

Investor's Corner

Tesla gets its best analysis from Morgan Stanley as ‘it’s all about to change’

He maintained its ‘Overweight’ rating and the $410 price target Morgan Stanley had on the stock.

Published

8 hours ago

on

August 7, 2025

By

(Credit: Tesla)

Tesla has gotten perhaps its best analysis from Morgan Stanley in quite some time, as the Wall Street firm claims that “it’s all about to change.”

That phrase could be used for both the company’s status and the world in general.

Analyst Adam Jonas said in a new note on Thursday to investors that Tesla could be one of the major winners in terms of the global transition from what it is now to what it will be.

He describes the global shift that will occur over the next few years:

“Have you interacted with a robot today? Have you even seen a robot today? No? Well, take a mental picture because it’s all about to change. When we meet someone who has never been in a Waymo or a Tesla Cybercab (which is most people), we frequently see a wince and a response such as ‘I’m not sure I’d feel comfortable getting in a car without a driver.’ We imagine going back in time to 1903 and asking people if they’d feel comfortable in an airplane.’”

Advertisement

The same technological revolutions that have occurred over the past 150 years will continue to occur again and again. We are on the verge of another, Jonas believes, as companies like Tesla are working on artificial intelligence tech, which includes changing the way we look at things like transportation and labor.

Jonas includes an interesting tidbit in his note about how humanoid robots could change wages, and how it could work into the advantage of Tesla, especially as it is developing its own Optimus robot:

“We estimate 1 humanoid robot at $5/hour can do the work of 2 humans at $25/hour, generating an NPV of approximately $200k/humanoid. 1 robot shaped car can potentially drive down cost/mile of a ride share vehicle to <$0.20 mile (1/10th human-driven ride-share).”

Jonas sees Tesla as a key player in how AI will impact things like manufacturing and various automotive industries, and he believes there is long-term potential for AI, robomobility, and even autonomous eVTOL platforms.

Tesla stock: Morgan Stanley says eVTOL is calling Elon Musk for new chapter

Advertisement

He maintained its ‘Overweight’ rating and the $410 price target Morgan Stanley had on the stock.

Continue Reading

Elon Musk

Tesla expands Robotaxi program in Austin to new riders

Tesla has been expanding both the rider group and the geofence in Austin slowly, making sure to prioritize safety and avoid any major events with the early rollout.

Published

9 hours ago

on

August 7, 2025

By

Credit: @TerrapinTerpene/X

Tesla is expanding its Robotaxi program in Austin, Texas, as several people have received invitations to participate and take rides.

Tesla first launched the Robotaxi platform on June 22. It invited a handful of people to participate in the first-ever public rides. We were lucky enough to get an invitation, and our permissions have been expanded in the Bay Area pilot program as well.

The group was small and consisted of big names in the Tesla community. It expanded and is continuing to offer these exclusive invitations to notable members of the Tesla community.

There have been fewer than five subsequent invitations after the first group’s were sent in late June:

Tesla has been expanding both the rider group and the geofence in Austin slowly, making sure to prioritize safety and avoid any major events with the early rollout.

Advertisement

Tesla’s new Robotaxi geofence shape is an FU by Elon Musk to the competition

“We are being very cautious. We do not want to take any chances, so we are going to go cautiously. But the service areas and the number of vehicles in operation will increase at a hyper-exponential rate,” CEO Elon Musk said during the Q2 Earnings Call.

Eventually, the Robotaxi platform will not require an invite, and it will operate without geofences. Musk believes Tesla can get there within three or six months, and plans to have at least half of the U.S. population with access to a Robotaxi by the end of the year:

Advertisement

“I think we will probably have autonomous ride-hailing in probably half the population of the U.S. by the end of the year. That’s at least our goal, subject to regulatory approvals. I think we will technically be able to do it. Assuming we have regulatory approvals, it’s probably addressing half the population of the U.S. by the end of the year.”

Tesla plans to have regulatory approval in Nevada, Arizona, and Florida sooner than in other states.

Continue Reading

Trending