QuantaBrain
Description
1 SUMMARY
The objective of QuantaBrain (Quantitative Analysis of Brain) is to improve the diagnosis process of psychiatric disorders using quantitative criteria by applying artificial intelligence algorithms to brain imaging data.
2 THE NEED
2.1 THE NEED OF A QUANTITATIVE PSYCHIATRY
It is estimated that more than 50% of people experience a mental disorder during their lifetime. In addition, 1 in 5 people in the United States suffer from a mental disorder every year and 1 in 25 lives have severe forms of schizophrenia, bipolar disorder and depression [1].
Unfortunately, since psychiatry is a mostly a qualitative science, based on the observation of behaviour and not on objective quantitative medical tests, the rate of misdiagnosis is very high. A study conducted in Canada shows that the rate of misdiagnosis is 65.9%, 85.8%, 92.7% and 97.8% for major depressive disorder, panic disorder, bipolar disorder and social anxiety disorder, respectively [2].
2.3 THE AUTISM SPECTRUM DISORDER
In addition to misdiagnosis, the lack of quantitative diagnostic tests causes delayed diagnosis, especially with regard to neurodevelopmental disorders. In fact, in the preverbal age it is very difficult to evaluate the behavioural symptoms of a child. For example, although many studies show that Autism Spectrum Disorder (ASD) is already present at birth [3], in the United States it is diagnosed on average around the age of 4.5. Furthermore, 30% of children do not receive a diagnosis of ASD before the age of 8 [4]. However, early diagnosis would be critical to allow children with ASD to promptly start behavioural treatments that are most effective if initiated within the critical period for the development of verbal and social skills, i.e. before the age of 2 [5].
3 THE OFFERING
3.1 IDEA BEHIND THE STARTUP
The idea behind the startup is to improve psychiatry by applying deep learning to quantitative tests such as brain imaging using EEG and / or MRI, in order to make the diagnosis process more accurate, objective and anticipate it.
Deep learning is able to identify complex biomarkers, that can change from individual to individual. For each person diagnosed, the characteristics that led the algorithm to make the diagnosis will be compiled in a report so that the doctor has the last word on the diagnosis and can use this information to start a personalized treatment.
As a first application, we focused on ASD which is an extremely heterogeneous and difficult to diagnose disorder and for which early diagnosis is crucial. In Italy, a study performed in 2018 on the incidence of ASD reports half that of the USA, suggesting that the disorder is still underdiagnosed [6]. Furthermore, an Italian study shows that 66.5% of subjects diagnosed in adulthood had received in the past one to eight psychiatric diagnoses different from ASD [7].
3.2 OUR ALGORITHM
Our algorithm uses a 6-minute functional magnetic resonance scan, acquired while the subject is at rest, and is able to diagnose ASD with a sensitivity and specificity of 88% and 85%. These figures are in line with the estimated sensitivity and specificity of experienced psychiatrists using ADOS (a questionnaire considered the golden standard for diagnosing ASD), which is 89-92% and 81-85% respectively [8] .
The results have been tested on more than 2000 subjects and on more than 40 different magnetic resonance machines and the performances are unchanged among all the age groups explored.
Our product uses the algorithm to generates a report that consists of a list of brain regions with atypical functioning that led to the diagnosis of the subject as DSA.
4 THE MAIN REVENUE GENERATION MECHANISM
QuantaBrain wants to offer subscriptions to hospitals for the use of a web platform that provides the diagnosis of a subject and a report on their brain characteristics upon uploading their anonymized MRI. In addition, as QuantaBrain collects MRI and statistics on brain regions involved in the disorder, these can be sold to pharmaceutical companies active in the field of personalized medicine, to accelerate drug research on ASD.
5 THE MARKET
5.1 THE NUMBERS THAT MATTER
The diagnosis of ASD clearly has an impact on the ASD population and on all subjects who in pre-verbal age have autistic traits for which it is advisable to make a diagnosis.
Our main markets will be the US and the EU, starting from the US, that are more open to this kind of technology.
ASD has an incidence of 1 in 44 [9] and some hospitals in the US that deal with early diagnosis (before age 2) indicate that 1 in 10 of those who are tested is then diagnosed as ASD. Considering that 1 out of 8 people have a neurodevelopmental disorder [10], the potential market if this application were to be extended to all subjects at risk becomes very large. Finally, as mentioned above, it is estimated that 50% of the population experiences some psychiatric condition in the course of their life.
From a commercial point of view, the web platform that will constitute our first product is aimed at centres specialized in the diagnosis of autism and in the USA these would be about 2,500 [11]. Furthermore, the data collected and analysed can be sold to pharmaceutical companies that carry out research and development on ASD, which according to the Autism Drug Trial Tracker [12] are about 117 in the world.
5.2 WHY AN HOSPITAL SHOULD MOVE TO OUR AUTOMATIC DIAGNOSIS
As for our first application, we know that the current diagnosis system is based on a series of psychiatric interviews with the child and his family that can last up to 14 hours. In the United States, the average cost of this service is $ 2,800. Instead, our system requires only 6 minutes of MRI and a split second to process the data, at a cost of $ 1,013 for a medical exam and about $ 600 for a hospital using our base subscription rate. So our system is faster and cheaper, as well as more objective and therefore suitable for early diagnosis.
Consequently, we expect hospitals adopting this diagnostic system to see:
• A reduction in diagnosis times and costs
• A greater influx of patients, as they will try to go to centers that allow for the safest and fastest diagnosis
• Greater patient satisfaction
• Correct early diagnosis, which implies that the resources previously used for all those subjects at risk, for whom the diagnosis was still uncertain, will be concentrated on the subjects who are really SLD.
6 THE UNIQUE FEATURES THAT YOU THINK YOUR IDEA HAS
6.1 COMPETITORS
There are no competitors that propose to carry out automatic and quantitative diagnosis of ASD, but there are companies that propose to diagnose autism with other approaches, or that use approaches similar to ours to diagnose other diseases.
Cognoa [13], a startup in the USA, offers an automatic diagnosis of autism based on the analysis of 3 inputs with artificial intelligence algorithms: a questionnaire made to families, one made to a doctor who visited the child and an evaluation carried out by an ASD expert who viewed a video of the child. Although this solution uses artificial intelligence, it is in fact based on the opinion of various people and therefore does not eliminate subjectivity and does not solve the problem of delay in diagnosis due to the difficulty in assessing the symptoms of ASD in the pre-verbal age. Nonetheless, Cognoa has raised more than 50 million dollars in Serie A [14]; which underlines how urgent the need for a change in the diagnostic system of ASD is perceived.
On the other hand, there are several startups that offer automatic diagnosis systems based on biomedical imaging, such as Arterys Inc [15], Digital Diagnostic [16] and Viz.AI [17]. However, none of them offer diagnostic systems for psychiatric diseases.
6.2 QUANTABRAIN DEVELOPED A SPECIFIC ALGORITHM FOR PSYCHIATRIC DIAGNOSIS
What most distinguishes QuantaBrain from other startups offering diagnostic services based on imaging data is that we diagnose psychiatric diseases.
Psychiatric diseases, unlike neurological diseases, do not have a known organic cause and although many molecular mechanisms underlying these disorders are known, inter-individual differences make it difficult to define a unique biomarker.
Today artificial intelligence is considered a key tool for understanding the complexity of these diseases. In the study behind this startup and the subject of the doctoral thesis of CEO, Elisa Ferrari, a new AI architecture is presented (for which a patent application has been submitted [18]) specifically designed for the diagnosis of psychiatric diseases.
The network has two crucial characteristics:
1) It uses adversarial training, i.e., training that seeks the right compromise between the that maximization of diagnostic performance and the minimization of the ability to recognize certain characteristics of the patient and / or image starting from the departure. This allowed us to build a network that always obtains the same performance on all possible subject groups: males and females, subjects with a different age or IQ, subjects who performed the scan following different instructions (e.g. open eyes / closed), data acquired with different scanners, etc. The adversarial training was crucial to solve well known problems in literature such as the sensitivity to the scanner model plaguing similar algorithms, or the fact that some algorithms were able to diagnose ASD subjects only if they were low-functioning (ie they had a low IQ).
2) It uses a dual-stream structure that starts from the functional magnetic resonance and from its optical flow. This solution has been used in the literature in video analysis to recognize actions, as the optical flow operation highlights the moving parts in a video. The rationale behind its use for the analysis of functional magnetic resonance images lies in the fact that if psychiatric diseases do not have an obvious structural cause, the anomaly must be sought in the functioning of the brain and therefore in the way in which the blood flows and oxygenates the parts of the brain involved in neuronal activity. The spatiotemporal pattern of blood in the brain is therefore a "movement" that could be recognized using this architecture.
This approach is sufficiently generic and can be used for the diagnosis of various psychiatric diseases.
In our first application on the diagnosis of ASD, in addition to verifying that the performance was sufficiently high, the brain regions that led the algorithm to give a positive diagnosis were then analysed. From this analysis, it emerged that a group of 19 out of 116 brain regions were particularly relevant for the diagnosis of ASD. A transcriptomics analysis was then conducted, based on the analysis of post-mortem brain samples and it emerged that those 19 regions are most influenced by genes that have been associated with ASD in literature.
This genetic validation of the algorithm performance provides us with further confirmation of its validity, leading us to believe that it can be applied to other diseases.
6.3 PSYCHIATRY WITH QUANTABRAIN
The innovation that QuantaBrain can bring to psychiatry by developing algorithms like the one just described and putting them on the market is articulated on several fronts:
• Families and patients will be comforted by a diagnosis that is safe, definitive, timely and cost-effective;
• Since the examination is quantitative, patients will be able to deliver it to all the specialists they wish to consult without having to repeat the diagnosis process;
• The psychiatrist will have less responsibility in the diagnosis phase, which will be short and automatic, and will therefore be able to devote more time to the delicate phase of treatment.
The indirect effects of this innovation will be that:
• Psychiatry will be more quantitative
• It will give patients and their families a greater awareness of the “physical” nature of psychiatric diseases, changing the way they are perceived and experienced.
• The diagnosis will be experienced less as a judgment and more as a medical examination.
• The diagnosis will be faster, less stressful and safer. In addition, the general practitioner himself could prescribe the exam, without having to go through the queues of specialist exams.
• MRI of the brain is a completely harmless and non-invasive examination. Since for our first application only 6 minutes of MRI are enough, we expect this instrument to be used as an early screening. Treatments could therefore become more ubiquitous and more precocious.
• The brain regions with atypical functioning identified by the algorithm can direct the patient towards a personalized program and accelerate the search for personalized drugs.
6.4 OUR POTENTIAL FOR TRANSFORMATION
Today QuantaBrain comes with a diagnostic algorithm for DSA, but as already mentioned, the method is generic and can be extended to other pathologies. A first medium-long term objective would be to extend this tool to other neurodevelopmental disorders, so as to constitute a universal screening tool. There are also several studies that use functional magnetic resonance imaging during pregnancy to study the functioning of the fetus' brain. A second long-term goal would be to use this screening tool even before birth, when some environmental risk factors during pregnancy [19] can still be corrected or mitigated [20]. Finally, we also aim to improve adult psychiatry in general and to allow all people to have a safe diagnosis and to be able to access the necessary treatment as soon as possible.
7 THE TEAM
7.1 BRIEF DESCRIPTION
The team has a solid scientific and multidisciplinary background. To date we are missing some members with a financial background, but we are supported by 42n, an entrepreneurship group that helps Italian startups to grow in the USA, that is providing us financial, economics and legal advices. In addition, Alberto Nobili and Hannah Teichmann, that are members of the team who worked for successful startups, are helping us with their experience and network of contacts.
7.2 INDIVIDUAL DESCRIPTIONS
Elisa Ferrari, CEO
In February 2022 Elisa Ferrari obtained her doctorate cum Laude in Data Science at the Scuola Nomale Superiore. Her doctoral thesis forms the idea behind QuantaBrain.
Elisa in 2017 won the PhD +, a competition launched by the University of Pisa for the best business ideas, obtaining a scholarship for an entrepreneurship acceleration course in California. After completing her doctoral studies, she was selected for a mentorship course by a group of Italian-Americans, called 42N, which is involved in helping Italian startups to enter the US market. Elisa is the inventor of 3 patents, all in the AI field, and now works full time at QuantaBrain.
Davide Bacciu, AI expert
Davide Bacciu is a professor of Computer Science at the University of Pisa and is considered an emerging figure in Italy in the field of artificial intelligence. He is vice-president of the AIxIA association (Italian association for artificial intelligence). He is one of the inventors of the patent behind this startup and has always been a great supporter of this initiative.
Alessandra Retico, MRI expert
Alessandra Retico has a PhD in Physics and is a researcher at the National Institute of Nuclear Physics, Section of Pisa. She is an expert in biomedical imaging, with particular experience in magnetic resonance. For several years now, one of her main research interests has been the diagnosis of Autism on the basis of neuroimaging data. She has collaborated with many hospitals dealing with Autism (eg Stella Maris Foundation, Gaslini, Bambino Gesù pediatric hospital, Meyer, etc.). She is also the inventor of the patent.
Alessandro Cellerino, Neurophysiology expert
Alessandro Cellerino is a professor of Neurobiology at the Scuola Normale Superiore. He studied the molecular processes and neuronal plasticity of the brain. He is the inventor of the patent at the base of the startup and has significantly contributed to the interpretation of the functional anomalies identified by our system in individuals with SLD.
Hannah Teichmann, expert in relations between clinicians and startup
Hannah Teichmann is a neuroscientist who founded Medical Microinstruments (MMI): a successful Italian company that recently closed a 75M round. She has been recognized as a MedTech Innovator top female founder in Italy (https: //www.thegoodintown. it / the-map-of-the-top-innovative-entrepreneurs-in-the-world /). Hannah believed in QuantaBrain from the very beginning and puts her many years of experience at the service of this startup.
Alberto Nobili, expert in bridging the Italian and US market
Alberto Nobili is an Italian who migrated to the USA to pursue his PhD in biotechnology at Harvard. Later he then founded his startup “Dynamic Cell Therapies” in Boston. Alberto is part of the “42n” group and of “Mass Medical Angels” and has always believed in this startup helping us to create a network in the USA.
8 REFERENCES
[1] https://www.cdc.gov/mentalhealth/learn/index.htm
[2] Vermani, Monica, Madalyn Marcus, and Martin A. Katzman. "Rates of detection of mood and anxiety disorders in primary care: a descriptive, cross-sectional study." The primary care companion for CNS disorders 13.2 (2011): 27211.
[3] https://kennethrobersonphd.com/can-you-develop-aspergers-syndrome-in-adulthood/
[4] https://www.cdc.gov/ncbddd/autism/addm-community-report/delay-to-accessing-services.html
[5] Gabbay-Dizdar, Nitzan, et al. "Early diagnosis of autism in the community is associated with marked improvement in social symptoms within 1–2 years." Autism 26.6 (2022): 1353-1363.
[6] Narzisi, A., et al. "Prevalence of Autism Spectrum Disorder in a large Italian catchment area: A school-based population study within the ASDEU project." Epidemiology and psychiatric sciences 29 (2020).
[7] Fusar-Poli, Laura, et al. "Missed diagnoses and misdiagnoses of adults with autism spectrum disorder." European archives of psychiatry and clinical neuroscience (2020): 1-12.
[8] Lebersfeld, Jenna B., et al. "Systematic review and meta-analysis of the clinical utility of the ADOS-2 and the ADI-R in diagnosing autism spectrum disorders in children." Journal of autism and developmental disorders 51.11 (2021): 4101-4114.
[9] https://www.cdc.gov/ncbddd/autism/addm.html
[10] Arora, Narendra K., et al. "Neurodevelopmental disorders in children aged 2–9 years: Population-based burden estimates across five regions in India." PLoS medicine 15.7 (2018): e1002615.
[11] Data from https://gapmap.stanford.edu/#/
[12] https://www.spectrumnews.org/news/introducing-spectrums-autism-drug-tracker/
[13] https://cognoa.com/
[14] https://www.crunchbase.com/organization/cognoa/company_financials
[15] https://www.arterys.com/
[16] https://www.digitaldiagnostics.com/
[17] https://www.viz.ai/
[18] Italian patent application filed on 24 March 2022, with no. 102022000005873
[19] van den Heuvel, Marion I., and Moriah E. Thomason. "Functional connectivity of the human brain in utero." Trends in cognitive sciences 20.12 (2016): 931-939.
[20] Lu, Jianping, et al. "Rethinking autism: The impact of maternal risk factors on autism development." American Journal of Translational Research 14.2 (2022): 1136.
Show More
The objective of QuantaBrain (Quantitative Analysis of Brain) is to improve the diagnosis process of psychiatric disorders using quantitative criteria by applying artificial intelligence algorithms to brain imaging data.
2 THE NEED
2.1 THE NEED OF A QUANTITATIVE PSYCHIATRY
It is estimated that more than 50% of people experience a mental disorder during their lifetime. In addition, 1 in 5 people in the United States suffer from a mental disorder every year and 1 in 25 lives have severe forms of schizophrenia, bipolar disorder and depression [1].
Unfortunately, since psychiatry is a mostly a qualitative science, based on the observation of behaviour and not on objective quantitative medical tests, the rate of misdiagnosis is very high. A study conducted in Canada shows that the rate of misdiagnosis is 65.9%, 85.8%, 92.7% and 97.8% for major depressive disorder, panic disorder, bipolar disorder and social anxiety disorder, respectively [2].
2.3 THE AUTISM SPECTRUM DISORDER
In addition to misdiagnosis, the lack of quantitative diagnostic tests causes delayed diagnosis, especially with regard to neurodevelopmental disorders. In fact, in the preverbal age it is very difficult to evaluate the behavioural symptoms of a child. For example, although many studies show that Autism Spectrum Disorder (ASD) is already present at birth [3], in the United States it is diagnosed on average around the age of 4.5. Furthermore, 30% of children do not receive a diagnosis of ASD before the age of 8 [4]. However, early diagnosis would be critical to allow children with ASD to promptly start behavioural treatments that are most effective if initiated within the critical period for the development of verbal and social skills, i.e. before the age of 2 [5].
3 THE OFFERING
3.1 IDEA BEHIND THE STARTUP
The idea behind the startup is to improve psychiatry by applying deep learning to quantitative tests such as brain imaging using EEG and / or MRI, in order to make the diagnosis process more accurate, objective and anticipate it.
Deep learning is able to identify complex biomarkers, that can change from individual to individual. For each person diagnosed, the characteristics that led the algorithm to make the diagnosis will be compiled in a report so that the doctor has the last word on the diagnosis and can use this information to start a personalized treatment.
As a first application, we focused on ASD which is an extremely heterogeneous and difficult to diagnose disorder and for which early diagnosis is crucial. In Italy, a study performed in 2018 on the incidence of ASD reports half that of the USA, suggesting that the disorder is still underdiagnosed [6]. Furthermore, an Italian study shows that 66.5% of subjects diagnosed in adulthood had received in the past one to eight psychiatric diagnoses different from ASD [7].
3.2 OUR ALGORITHM
Our algorithm uses a 6-minute functional magnetic resonance scan, acquired while the subject is at rest, and is able to diagnose ASD with a sensitivity and specificity of 88% and 85%. These figures are in line with the estimated sensitivity and specificity of experienced psychiatrists using ADOS (a questionnaire considered the golden standard for diagnosing ASD), which is 89-92% and 81-85% respectively [8] .
The results have been tested on more than 2000 subjects and on more than 40 different magnetic resonance machines and the performances are unchanged among all the age groups explored.
Our product uses the algorithm to generates a report that consists of a list of brain regions with atypical functioning that led to the diagnosis of the subject as DSA.
4 THE MAIN REVENUE GENERATION MECHANISM
QuantaBrain wants to offer subscriptions to hospitals for the use of a web platform that provides the diagnosis of a subject and a report on their brain characteristics upon uploading their anonymized MRI. In addition, as QuantaBrain collects MRI and statistics on brain regions involved in the disorder, these can be sold to pharmaceutical companies active in the field of personalized medicine, to accelerate drug research on ASD.
5 THE MARKET
5.1 THE NUMBERS THAT MATTER
The diagnosis of ASD clearly has an impact on the ASD population and on all subjects who in pre-verbal age have autistic traits for which it is advisable to make a diagnosis.
Our main markets will be the US and the EU, starting from the US, that are more open to this kind of technology.
ASD has an incidence of 1 in 44 [9] and some hospitals in the US that deal with early diagnosis (before age 2) indicate that 1 in 10 of those who are tested is then diagnosed as ASD. Considering that 1 out of 8 people have a neurodevelopmental disorder [10], the potential market if this application were to be extended to all subjects at risk becomes very large. Finally, as mentioned above, it is estimated that 50% of the population experiences some psychiatric condition in the course of their life.
From a commercial point of view, the web platform that will constitute our first product is aimed at centres specialized in the diagnosis of autism and in the USA these would be about 2,500 [11]. Furthermore, the data collected and analysed can be sold to pharmaceutical companies that carry out research and development on ASD, which according to the Autism Drug Trial Tracker [12] are about 117 in the world.
5.2 WHY AN HOSPITAL SHOULD MOVE TO OUR AUTOMATIC DIAGNOSIS
As for our first application, we know that the current diagnosis system is based on a series of psychiatric interviews with the child and his family that can last up to 14 hours. In the United States, the average cost of this service is $ 2,800. Instead, our system requires only 6 minutes of MRI and a split second to process the data, at a cost of $ 1,013 for a medical exam and about $ 600 for a hospital using our base subscription rate. So our system is faster and cheaper, as well as more objective and therefore suitable for early diagnosis.
Consequently, we expect hospitals adopting this diagnostic system to see:
• A reduction in diagnosis times and costs
• A greater influx of patients, as they will try to go to centers that allow for the safest and fastest diagnosis
• Greater patient satisfaction
• Correct early diagnosis, which implies that the resources previously used for all those subjects at risk, for whom the diagnosis was still uncertain, will be concentrated on the subjects who are really SLD.
6 THE UNIQUE FEATURES THAT YOU THINK YOUR IDEA HAS
6.1 COMPETITORS
There are no competitors that propose to carry out automatic and quantitative diagnosis of ASD, but there are companies that propose to diagnose autism with other approaches, or that use approaches similar to ours to diagnose other diseases.
Cognoa [13], a startup in the USA, offers an automatic diagnosis of autism based on the analysis of 3 inputs with artificial intelligence algorithms: a questionnaire made to families, one made to a doctor who visited the child and an evaluation carried out by an ASD expert who viewed a video of the child. Although this solution uses artificial intelligence, it is in fact based on the opinion of various people and therefore does not eliminate subjectivity and does not solve the problem of delay in diagnosis due to the difficulty in assessing the symptoms of ASD in the pre-verbal age. Nonetheless, Cognoa has raised more than 50 million dollars in Serie A [14]; which underlines how urgent the need for a change in the diagnostic system of ASD is perceived.
On the other hand, there are several startups that offer automatic diagnosis systems based on biomedical imaging, such as Arterys Inc [15], Digital Diagnostic [16] and Viz.AI [17]. However, none of them offer diagnostic systems for psychiatric diseases.
6.2 QUANTABRAIN DEVELOPED A SPECIFIC ALGORITHM FOR PSYCHIATRIC DIAGNOSIS
What most distinguishes QuantaBrain from other startups offering diagnostic services based on imaging data is that we diagnose psychiatric diseases.
Psychiatric diseases, unlike neurological diseases, do not have a known organic cause and although many molecular mechanisms underlying these disorders are known, inter-individual differences make it difficult to define a unique biomarker.
Today artificial intelligence is considered a key tool for understanding the complexity of these diseases. In the study behind this startup and the subject of the doctoral thesis of CEO, Elisa Ferrari, a new AI architecture is presented (for which a patent application has been submitted [18]) specifically designed for the diagnosis of psychiatric diseases.
The network has two crucial characteristics:
1) It uses adversarial training, i.e., training that seeks the right compromise between the that maximization of diagnostic performance and the minimization of the ability to recognize certain characteristics of the patient and / or image starting from the departure. This allowed us to build a network that always obtains the same performance on all possible subject groups: males and females, subjects with a different age or IQ, subjects who performed the scan following different instructions (e.g. open eyes / closed), data acquired with different scanners, etc. The adversarial training was crucial to solve well known problems in literature such as the sensitivity to the scanner model plaguing similar algorithms, or the fact that some algorithms were able to diagnose ASD subjects only if they were low-functioning (ie they had a low IQ).
2) It uses a dual-stream structure that starts from the functional magnetic resonance and from its optical flow. This solution has been used in the literature in video analysis to recognize actions, as the optical flow operation highlights the moving parts in a video. The rationale behind its use for the analysis of functional magnetic resonance images lies in the fact that if psychiatric diseases do not have an obvious structural cause, the anomaly must be sought in the functioning of the brain and therefore in the way in which the blood flows and oxygenates the parts of the brain involved in neuronal activity. The spatiotemporal pattern of blood in the brain is therefore a "movement" that could be recognized using this architecture.
This approach is sufficiently generic and can be used for the diagnosis of various psychiatric diseases.
In our first application on the diagnosis of ASD, in addition to verifying that the performance was sufficiently high, the brain regions that led the algorithm to give a positive diagnosis were then analysed. From this analysis, it emerged that a group of 19 out of 116 brain regions were particularly relevant for the diagnosis of ASD. A transcriptomics analysis was then conducted, based on the analysis of post-mortem brain samples and it emerged that those 19 regions are most influenced by genes that have been associated with ASD in literature.
This genetic validation of the algorithm performance provides us with further confirmation of its validity, leading us to believe that it can be applied to other diseases.
6.3 PSYCHIATRY WITH QUANTABRAIN
The innovation that QuantaBrain can bring to psychiatry by developing algorithms like the one just described and putting them on the market is articulated on several fronts:
• Families and patients will be comforted by a diagnosis that is safe, definitive, timely and cost-effective;
• Since the examination is quantitative, patients will be able to deliver it to all the specialists they wish to consult without having to repeat the diagnosis process;
• The psychiatrist will have less responsibility in the diagnosis phase, which will be short and automatic, and will therefore be able to devote more time to the delicate phase of treatment.
The indirect effects of this innovation will be that:
• Psychiatry will be more quantitative
• It will give patients and their families a greater awareness of the “physical” nature of psychiatric diseases, changing the way they are perceived and experienced.
• The diagnosis will be experienced less as a judgment and more as a medical examination.
• The diagnosis will be faster, less stressful and safer. In addition, the general practitioner himself could prescribe the exam, without having to go through the queues of specialist exams.
• MRI of the brain is a completely harmless and non-invasive examination. Since for our first application only 6 minutes of MRI are enough, we expect this instrument to be used as an early screening. Treatments could therefore become more ubiquitous and more precocious.
• The brain regions with atypical functioning identified by the algorithm can direct the patient towards a personalized program and accelerate the search for personalized drugs.
6.4 OUR POTENTIAL FOR TRANSFORMATION
Today QuantaBrain comes with a diagnostic algorithm for DSA, but as already mentioned, the method is generic and can be extended to other pathologies. A first medium-long term objective would be to extend this tool to other neurodevelopmental disorders, so as to constitute a universal screening tool. There are also several studies that use functional magnetic resonance imaging during pregnancy to study the functioning of the fetus' brain. A second long-term goal would be to use this screening tool even before birth, when some environmental risk factors during pregnancy [19] can still be corrected or mitigated [20]. Finally, we also aim to improve adult psychiatry in general and to allow all people to have a safe diagnosis and to be able to access the necessary treatment as soon as possible.
7 THE TEAM
7.1 BRIEF DESCRIPTION
The team has a solid scientific and multidisciplinary background. To date we are missing some members with a financial background, but we are supported by 42n, an entrepreneurship group that helps Italian startups to grow in the USA, that is providing us financial, economics and legal advices. In addition, Alberto Nobili and Hannah Teichmann, that are members of the team who worked for successful startups, are helping us with their experience and network of contacts.
7.2 INDIVIDUAL DESCRIPTIONS
Elisa Ferrari, CEO
In February 2022 Elisa Ferrari obtained her doctorate cum Laude in Data Science at the Scuola Nomale Superiore. Her doctoral thesis forms the idea behind QuantaBrain.
Elisa in 2017 won the PhD +, a competition launched by the University of Pisa for the best business ideas, obtaining a scholarship for an entrepreneurship acceleration course in California. After completing her doctoral studies, she was selected for a mentorship course by a group of Italian-Americans, called 42N, which is involved in helping Italian startups to enter the US market. Elisa is the inventor of 3 patents, all in the AI field, and now works full time at QuantaBrain.
Davide Bacciu, AI expert
Davide Bacciu is a professor of Computer Science at the University of Pisa and is considered an emerging figure in Italy in the field of artificial intelligence. He is vice-president of the AIxIA association (Italian association for artificial intelligence). He is one of the inventors of the patent behind this startup and has always been a great supporter of this initiative.
Alessandra Retico, MRI expert
Alessandra Retico has a PhD in Physics and is a researcher at the National Institute of Nuclear Physics, Section of Pisa. She is an expert in biomedical imaging, with particular experience in magnetic resonance. For several years now, one of her main research interests has been the diagnosis of Autism on the basis of neuroimaging data. She has collaborated with many hospitals dealing with Autism (eg Stella Maris Foundation, Gaslini, Bambino Gesù pediatric hospital, Meyer, etc.). She is also the inventor of the patent.
Alessandro Cellerino, Neurophysiology expert
Alessandro Cellerino is a professor of Neurobiology at the Scuola Normale Superiore. He studied the molecular processes and neuronal plasticity of the brain. He is the inventor of the patent at the base of the startup and has significantly contributed to the interpretation of the functional anomalies identified by our system in individuals with SLD.
Hannah Teichmann, expert in relations between clinicians and startup
Hannah Teichmann is a neuroscientist who founded Medical Microinstruments (MMI): a successful Italian company that recently closed a 75M round. She has been recognized as a MedTech Innovator top female founder in Italy (https: //www.thegoodintown. it / the-map-of-the-top-innovative-entrepreneurs-in-the-world /). Hannah believed in QuantaBrain from the very beginning and puts her many years of experience at the service of this startup.
Alberto Nobili, expert in bridging the Italian and US market
Alberto Nobili is an Italian who migrated to the USA to pursue his PhD in biotechnology at Harvard. Later he then founded his startup “Dynamic Cell Therapies” in Boston. Alberto is part of the “42n” group and of “Mass Medical Angels” and has always believed in this startup helping us to create a network in the USA.
8 REFERENCES
[1] https://www.cdc.gov/mentalhealth/learn/index.htm
[2] Vermani, Monica, Madalyn Marcus, and Martin A. Katzman. "Rates of detection of mood and anxiety disorders in primary care: a descriptive, cross-sectional study." The primary care companion for CNS disorders 13.2 (2011): 27211.
[3] https://kennethrobersonphd.com/can-you-develop-aspergers-syndrome-in-adulthood/
[4] https://www.cdc.gov/ncbddd/autism/addm-community-report/delay-to-accessing-services.html
[5] Gabbay-Dizdar, Nitzan, et al. "Early diagnosis of autism in the community is associated with marked improvement in social symptoms within 1–2 years." Autism 26.6 (2022): 1353-1363.
[6] Narzisi, A., et al. "Prevalence of Autism Spectrum Disorder in a large Italian catchment area: A school-based population study within the ASDEU project." Epidemiology and psychiatric sciences 29 (2020).
[7] Fusar-Poli, Laura, et al. "Missed diagnoses and misdiagnoses of adults with autism spectrum disorder." European archives of psychiatry and clinical neuroscience (2020): 1-12.
[8] Lebersfeld, Jenna B., et al. "Systematic review and meta-analysis of the clinical utility of the ADOS-2 and the ADI-R in diagnosing autism spectrum disorders in children." Journal of autism and developmental disorders 51.11 (2021): 4101-4114.
[9] https://www.cdc.gov/ncbddd/autism/addm.html
[10] Arora, Narendra K., et al. "Neurodevelopmental disorders in children aged 2–9 years: Population-based burden estimates across five regions in India." PLoS medicine 15.7 (2018): e1002615.
[11] Data from https://gapmap.stanford.edu/#/
[12] https://www.spectrumnews.org/news/introducing-spectrums-autism-drug-tracker/
[13] https://cognoa.com/
[14] https://www.crunchbase.com/organization/cognoa/company_financials
[15] https://www.arterys.com/
[16] https://www.digitaldiagnostics.com/
[17] https://www.viz.ai/
[18] Italian patent application filed on 24 March 2022, with no. 102022000005873
[19] van den Heuvel, Marion I., and Moriah E. Thomason. "Functional connectivity of the human brain in utero." Trends in cognitive sciences 20.12 (2016): 931-939.
[20] Lu, Jianping, et al. "Rethinking autism: The impact of maternal risk factors on autism development." American Journal of Translational Research 14.2 (2022): 1136.
Sustainable Development Goals
3. Good Health And Well-Being
Our startup helps to improve and anticipate autism diagnosis, so it is supported by this SDG because: 1) it is related to child health, which is a major concern for this SDG 2) since 2015 mental health is included in this SDG [1] REFERENCES [1] Votruba, Nicole, Graham Thornicroft, and FundaMentalSDG Steering Group. "Sustainable development goals and mental health: learnings from the contribution of the FundaMentalSDG global initiative." Global Mental Health 3 (2016).
Save to favorites
Notice: Undefined variable: post_id in /var/www/vhosts/hei-prometheus.eu/httpdocs/entrepreneurship-acceleration-platform/wp-content/themes/understrap-child-1/loop-templates/content-businessidea.php on line 293
Notice: Undefined variable: post_id in /var/www/vhosts/hei-prometheus.eu/httpdocs/entrepreneurship-acceleration-platform/wp-content/themes/understrap-child-1/loop-templates/content-businessidea.php on line 293
Members
1
Technologies
AI
Cloud Computing
Data Analytics – Big Data
Innovative Software
Product Type - System
B2B Services
Sector
Medicine and Health
Belongs to a competition
No
Thematic Area
No
Categories
No
Type of need
Collaboration - We want help in setting up our business
We can offer to these hospitals free or discounted subscriptions. In the meantime, we can co-author scienfic publications on this technology.
Collaboration - We want a new member in our team
We can offer him/her the CMO (Chief Medical Officier) position along with equity and possibly a salary (depending on the degree of his/her involvment in the startup).
Investors - We need investors to support our idea
The investment can be done in exchange of direct equity or through the use of the "simple agreements for future equity".