Source study found

Story checked

Cambridge scientists create a $10 chip that could put a medical lab on your wrist | ScienceDaily (opens in a new tab)

sciencedaily.com · 2026-10-22

Short answerEvidenceSource

Short answer

Mixed

Mixed.

The claims we could check match the study, but some claims were not covered by the evidence reviewed.

2 supported
3 not covered

Checked against the study summary. The full text wasn't available, so some details couldn't be settled either way.

Share this check

Follow the evidence trail

The story

Cambridge scientists create a $10 chip that could put a medical lab on your wrist | ScienceDaily

sciencedaily.com · 2026-10-22

The story’s checkable claims.

Read the original story

NewsLink checks it

Mixed

Every claim we could check holds up. Two of five claims match the study. This overall rating is based only on the claims we could check. Three claims the study doesn't address.

2 supported
3 not covered

Open claim evidence

The source study

Optical convolutional spectrometer

Nature Photonics · 2026Open access

Open source record Open in Tessa

Then inspect each claim

Evidence layer

Claim by claim

Each claim gets a verdict. Expand it to see the evidence directly below.

Reading mode

Scan verdicts. Open evidence only when needed.

Browse by verdict

5 claims in this story

Showing all 5 claimsChoose a verdict to focus the list.

Claim 1 of 5Not coveredThe researchers say the chip is a convolutional spectrometer that uses repeating optical components and fast Fourier transforms to recover spectra with high precision, low computational overhead, and broad near-infrared coverage.View evidenceHide evidence

As stated1200-1700 nm

Why this verdict

The abstract supports that the device is a convolutional spectrometer based on the convolution theorem, with wide near-infrared bandwidth, high precision, computational simplicity, and subsecond sampling/processing. However, the supplied abstract-level profile does not verify the more specific story details about repeating optical components, fast Fourier transforms, or the exact 1200–1700 nm wavelength range. Those may be in the full paper, but they are not verifiable from the provided abstract-depth evidence.

Study evidence

The paper presents a convolutional spectrometer design claiming a centimetre-scale fully packaged footprint and low estimated unit cost (~US$10).centimetre-scale footprint; ~US$10

“Here we introduce a new class of spectrometer, which uses the convolution theorem as its unique mathematical foundation.”

Claim 2 of 5Not coveredIn reported experiments, the spectrometer measured water-based and organic solutions with about 0.01% accuracy and worked reliably across temperatures from -20°C to 80°C.View evidenceHide evidence

As statedapproximately 0.01%

Why this verdict

The abstract supports that aqueous and organic solution concentrations were quantified and says detection accuracy surpassed commercial benchtop spectrometers. It does not provide the story's specific approximately 0.01% accuracy figure, nor does it mention reliable operation across -20°C to 80°C. These details are therefore not verifiable at the supplied abstract depth.

Study evidence

Quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers.

“and quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers.”

Claim 3 of 5Not coveredThe article says the device accurately measured several human biomarkers, including skin moisture, blood alcohol, blood lactate, and blood glucose, and demonstrated noninvasive glucose tracking over an extended period in one participant.View evidenceHide evidence

As statedone participant

Why this verdict

The abstract supports non-invasive sensing of skin moisture, blood alcohol, blood lactate, and blood glucose, with reported mean absolute errors. However, it does not report the participant count, an extended-period glucose-tracking demonstration, or that this was in one participant. The biomarker-sensing portion is supported, but the extended one-participant glucose-tracking detail is not verifiable from the abstract-level profile.

Study evidence

Skin moisture estimated non-invasively from near-infrared spectra with mean absolute error reported as 2.49%.MAE = 2.49%

“We also realize the non-invasive sensing of human biomarkers, such as skin moisture (mean absolute error = 2.49%), blood alcohol (1.70 mg dl-1), blood lactate (0.81 mmol l-1) and blood glucose (0.36 mmol l-1)”

Claim 4 of 5SupportedCambridge researchers have created a $10 optical sensor that delivers laboratory-level precision in a device small enough for wearable technology.View evidenceHide evidence

As stated$10

Why this verdict

At abstract depth, the paper supports the main elements: a miniaturized convolutional spectrometer with an estimated low cost of about US$10, a centimetre-scale packaged footprint, wide near-infrared bandwidth, subsecond operation, and claims of high precision including solution quantification said to surpass commercial benchtop spectrometers. The headline phrasing 'delivers laboratory-level precision' is promotional and broad, but it is broadly aligned with the abstract's own lab/benchtop-performance framing.

Study evidence

The paper presents a convolutional spectrometer design claiming a centimetre-scale fully packaged footprint and low estimated unit cost (~US$10).centimetre-scale footprint; ~US$10

“Here we introduce a new class of spectrometer, which uses the convolution theorem as its unique mathematical foundation.”

Study evidence

Quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers.

“and quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers.”

Claim 5 of 5SupportedThe device identified foods, plastics, and medicines with 100% accuracy in tests and could monitor blood glucose, alcohol, and other health markers without invasive procedures.View evidenceHide evidence

As stated100% accuracy

Why this verdict

The abstract reports 100% classification success for diverse solid samples including plastics, pharmaceuticals, coffee, flour and tea, which maps to plastics, medicines, and foods. It also reports non-invasive sensing of human biomarkers including blood glucose, blood alcohol, blood lactate, and skin moisture. The claim is framed as demonstrated ability/potential rather than established clinical monitoring, so it is supported at abstract depth, though the abstract lacks validation details.

Study evidence

The convolutional spectrometer classified a set of diverse solid samples (plastics, pharmaceuticals, coffee, flour, tea) with a reported 100% success rate.100% accuracy

“Specifically, our spectrometer system classifies diverse solid samples, including plastics, pharmaceuticals, coffee, flour and tea, with 100% success rate”

Study evidence

Skin moisture estimated non-invasively from near-infrared spectra with mean absolute error reported as 2.49%.MAE = 2.49%

“We also realize the non-invasive sensing of human biomarkers, such as skin moisture (mean absolute error = 2.49%), blood alcohol (1.70 mg dl-1), blood lactate (0.81 mmol l-1) and blood glucose (0.36 mmol l-1)”

Then look for missing context

Context layer

What the story left out

Important study details the story did not include.

Reported biomarker error metrics: skin moisture MAE 2.49%, blood alcohol MAE 1.70 mg dl^-1, blood lactate MAE 0.81 mmol l^-1, and blood glucose MAE 0.36 mmol l^-1.
The story says the device accurately measured biomarkers but does not include the paper's reported quantitative error metrics, which are material to interpreting the strength of the biomarker claims.
From observational (human in vivo)

6 things the story did carry across

The paper's primary contribution is the introduction of a miniaturized optical convolutional spectrometer based on the convolution theorem.
Device-level specifications: approximately US$10 cost, centimetre-scale packaged footprint, wide near-infrared bandwidth of about 2,400 cm^-1 / 500 nm, and subsecond sampling and processing.
Solid-sample classification demonstration: plastics, pharmaceuticals, coffee, flour, and tea classified with a reported 100% success rate.
Solution quantification demonstration: aqueous and organic solution concentrations quantified with detection accuracy claimed to surpass commercial benchtop spectrometers.
Human non-invasive biomarker sensing: skin moisture, blood alcohol, blood lactate, and blood glucose measured with reported prediction errors.
Generalizability, robustness across conditions, and manufacturability are not established from the abstract-level evidence.

Then read the study layer

Study layer

Study at a glance

Scan the study first. Expand only the parts you want to inspect.

Pieces of work

Evidence read

study summary

Lead result

other

1Lead resultotherIntroduce and describe an optical “convolutional spectrometer” whose operating principle is based on the convolution theorem, emphasizing miniaturized, low-cost, wide-bandwidth near-infrared performance with subsecond acquisition/processing.Convolutional spectrometer (hardware + computational reconstruction)Expand

In plain English

Paper introduces a new class of miniaturized "convolutional spectrometer" whose operating principle and reconstruction algorithm are based on the convolution theorem. The authors report a centimetre-scale fully packaged device footprint, low unit cost (~US$10), a wide near-infrared bandwidth (2,400 cm^-1 ≈ 500 nm), and subsecond sampling and processing. Demonstrations in the abstract include 100% classification accuracy on diverse solid samples, concentration quantification of solutions with accuracy said to surpass commercial benchtop spectrometers, and non‑invasive sensing of human biomarkers (skin moisture, blood alcohol, blood lactate, blood glucose) with reported mean absolute errors.

Key findings

The paper presents a convolutional spectrometer design claiming a centimetre-scale fully packaged footprint and low estimated unit cost (~US$10).centimetre-scale footprint; ~US$10
The system provides a wide near-infrared bandwidth of 2,400 cm^-1 (approximately 500 nm).2,400 cm^-1 (~500 nm)

“Here we introduce a new class of spectrometer, which uses the convolution theorem as its unique mathematical foundation.”

What this piece can’t prove

3 further details could not be confirmed from the summary.

2otherDemonstrate analytical performance on non-human materials: classify diverse solid samples (plastics, pharmaceuticals, coffee, flour, tea) and quantify concentrations of aqueous/organic solutions, benchmarked against commercial benchtop spectrometers.NIR spectroscopy + supervised classification for solid sample identificationExpand

In plain English

The authors report empirical validation of the convolutional spectrometer for non‑human solid samples: using near‑infrared measurements plus supervised classification, the system reportedly classifies a set of diverse solids (plastics, pharmaceuticals, coffee, flour, tea) with a claimed 100% success rate and operates with subsecond sampling and processing.

Key findings

The convolutional spectrometer classified a set of diverse solid samples (plastics, pharmaceuticals, coffee, flour, tea) with a reported 100% success rate.100% accuracy

“Specifically, our spectrometer system classifies diverse solid samples, including plastics, pharmaceuticals, coffee, flour and tea, with 100% success rate”

What this piece can’t prove

Summary is based solely on abstract statements; the abstract does not report sample sizes, dataset composition, or the statistical validation approach used for the classification claim.
Abstract does not provide per‑class performance, confusion matrices, or comparative benchmarking results specific to the solid‑sample classification task.

1 further detail could not be confirmed from the summary.

3otherDemonstrate analytical performance on non-human materials: classify diverse solid samples (plastics, pharmaceuticals, coffee, flour, tea) and quantify concentrations of aqueous/organic solutions, benchmarked against commercial benchtop spectrometers.Near-infrared quantitative spectroscopy and benchmarkingExpand

In plain English

The abstract reports that the presented convolutional spectrometer quantifies concentrations of aqueous and organic solutions with detection accuracy that surpasses commercial benchtop spectrometers. The device is described as a centimetre-scale, low-cost (~US$10) near-infrared instrument with ~2,400 cm-1 bandwidth and subsecond sampling and processing time.

Key findings

Quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers.

“and quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers.”

What this piece can’t prove

3 further details could not be confirmed from the summary.

4human in vivoDemonstrate non-invasive sensing of human biomarkers (skin moisture, blood alcohol, lactate, glucose) using near-infrared spectra, reporting prediction errors.observational (human in vivo)Expand

In plain English

The paper reports non-invasive near-infrared spectroscopic sensing of human biomarkers using the convolutional spectrometer, with quantitative prediction errors reported for skin moisture and blood analytes (blood alcohol, blood lactate, blood glucose).

Key findings

Skin moisture estimated non-invasively from near-infrared spectra with mean absolute error reported as 2.49%.MAE = 2.49%
Blood alcohol concentration estimated non-invasively with mean absolute error reported as 1.70 mg dl-1.MAE = 1.70 mg dl-1

“We also realize the non-invasive sensing of human biomarkers, such as skin moisture (mean absolute error = 2.49%), blood alcohol (1.70 mg dl-1), blood lactate (0.81 mmol l-1) and blood glucose (0.36 mmol l-1)”

What this piece can’t prove

Unclear whether reported MAEs derive from independent test sets, cross-validation, or training data performance.
Abstract does not report limits of detection, dynamic ranges, or performance across different skin types, physiological conditions, or potential confounders.

2 further details could not be confirmed from the summary.

Finally, the search trail

Method layer