Frequently Asked Questions

Last updated: March 2017.

These are quick, informal answers to frequently asked questions. For more detailed information with references to the scientific literature, see our formal research proposal. You may also find our Google talk slides to be informative.

What is your privacy policy concerning volunteer data?

See this page.

What is g?

No one knows precisely what intelligence is, and even experts disagree as to how it should be defined. However, it has been known for over a century that performance on different cognitive tests is positively correlated: for example, someone who is good at math puzzles is also more likely to have an above average vocabulary. Given a battery of tests and their correlation matrix, one can use probability theory to define a single parameter that, in a sense, optimally compresses the information from administering them all.

In practice, a wide range of intuitively sensible test batteries and functions of their score vectors yield very similar estimates of this parameter. As a result, psychologists consider these functions of test batteries to all be reasonable estimators of a parameter called the General Factor of Intelligence, or g for short.

Why is g important?

The human brain is extremely complex, a unique product of millions of years of evolution. Our genetic code is the "blueprint" from which this object is constructed. The genetics of cognition inform subjects as diverse as psychology, anthropology, neuroscience and molecular biology. g may be a rough guide to the overall goodness of function of the brain.

In addition, g has significant correlations with health outcomes. It has been demonstrated to be amongst the very best predictors of cardiovascular disease, resistance to dementia, ability to quit smoking, and even simple longevity. The brain is an organ like any other, and it affects human health and well-being profoundly.

How precisely can g be defined?

Note that the definition of g is periodically refined. However, it is undoubtedly a robust phenomenon, and a promising metric upon which to base our study. Some important properties of g are:

  1. stability (scores tend not to vary significantly after adolescence),
  2. heritability (twin studies and adoption studies powerfully suggest that much of the variance in g is dependent upon genetics), and
  3. predictive power (g scores are correlated with academic and job performance, income, longevity, etc., even after controlling for other variables such as social class, ethnic background, and resource access).

How heritable is g?

Twin and adoption studies suggest that at least 50%, and perhaps as much as 80%, of the variance in g scores is due to genetic causes. Note that heritability is defined over a specific distribution of environments. For the 50-80% range mentioned above, the environments in question are those found in families typically allowed to adopt foster children. Severe environmental deprivation may reduce heritability, as with most traits, but the predictive value upon g of genetics has long been well established.

Why collect this volunteer cohort?

Our volunteers have done so to help push the boundaries of scientific knowledge. High g individuals are the recipients of a special gift. By studying their DNA, we may be able to determine which loci in the human genome affect cognitive development, and by extension, human health and well-being.

Will my personal data and genomic data (gata) be safe?

Absolutely yes. We will never release your identity without multiple confirmations of your explicit approval, and our study is designed with particular care for adherence to strict security requirements.

  1. Genetic data (gata) is stored separately from personally identifying information, to reduce the probability and the effect of any security compromise.
  2. Correlation between gata and personally identifying information is only through masked ID's, the keys to which are stored offline.
  3. Etc.—see our privacy statement for more details. We take privacy and data security very seriously.

Note that hundreds of thousands of people have been research participants in genome-wide association studies of height, body mass index, diabetes, and other traits. To our knowledge, no individual has been harmed by participation in such a study. The track record does show that widely adopted guidelines for data security and the protection of research participants reduce risk to a minimum.

Also, BGI is not a government institution, and no data will be turned over to any government entity, particularly not the Chinese government. BGI has throughout its history collaborated extensively with researchers based in multiple countries, at all times adhering to standard protocols for data security and the protection of research participants. The most well-known of these collaborations are the Human Genome Project, the International HapMap Project, and the 1000 Genomes Project. One recent international collaboration involving BGI is this study examining the exome sequences of 200 Danish nationals. This collaboration included scientists from UC Berkeley, the University of Copenhagen in Denmark, and several other institutions.

Moreover, all members of the BGI Cognitive Genomics Lab with access to personal information entered on this website will abide by the terms of our privacy statement.

If these provisions are not sufficient to put you at ease, please let us know your concerns at Again: we take privacy and data security very seriously.

Are you still receiving new volunteers?

We currently are not seeking new volunteers.

How did one qualify for the g study?

We have sought participants with high cognitive ability. Volunteers could qualify for the study by having obtained a high SAT/ACT/GRE score, or having performed well in academic competitions such as the Math, Physics, or Informatics Olympiads, the William Lowell Putnam Mathematical Competition, TopCoder, etc. They could also qualify via exceptional academic credentials or technical accomplishments.

Are your screening criteria biased toward math ability?

Yes. Our study is international, and "mathematical ability" is the largest coherent component of g for which (i) enough people have already been tested and (ii) cross-country comparisons are practical. Please keep in mind that we must study people from a variety of countries and cultures.

We do not claim that our study design is capable of identifying all g-associated alleles given enough participants, let alone all loci linked with other components of intelligence; or that g is a perfect measurement of intelligence, brain health, etc. We simply wish to help start the process of discovery, and believe that this is a good place to begin. Keep in mind the difficulty of this line of inquiry: previous intelligence GWAS, even while confirming additive g heritability of at least 0.45 (see Davies et al.'s recent paper), have not been successful in finding replicable associations. (Update, 1 June 2013: Rietveld et al.'s large educational attainment GWAS may have cleared this hurdle.)

It is reasonable to interpret the study of g as being of something narrower than "intelligence", yet with important correlations for human health and well-being nonetheless.

What is prosopagnosia?

Prosopagnosia, also called face blindness, is an impairment in the recognition of faces. It is often accompanied by other types of recognition impairments (place recognition, car recognition, facial expression of emotion, etc.) though sometimes it appears to be restricted to facial identity. Not surprisingly, prosopagnosia can create serious social problems. Prosopagnosics often have difficulty recognizing family members, close friends, and even themselves. They often use alternative routes to recognition, but these routes are not as effective as recognition via the face.

I have another question...

First, check if our research proposal answers it. If not, feel free to ask us at; we'll try to reply in a reasonable amount of time, and especially good questions may be added to this page.