Dec. 9, 1993
Vol. 13, No. 8

current issue
archive / search

    Wong's work in statistics transcends traditional boundari

    Research in statistics frequently extends beyond traditional departmental boundaries, and the work of Wing Wong, Professor in Statistics, is an excellent example of that pattern.

    Although he is a theoretical statistician, Wong often collaborates with scientists in other departments to solve practical problems that involve the theoretical methods he has developed.

    In his most recent interdisciplinary projects, Wong has collaborated with geneticists, radiologists and chemists to solve problems that require statistical estimation -- such problems as how to convert a detector's signals into a multidimensional "picture" of a human brain and how to predict the way a protein is going to fold in three dimensions.

    Such collaboration with other scientists was among the qualities cited by the Committee of Presidents of Statistical Societies in presenting its Presidents' Award to Wong. The award, an international prize given annually since 1981 to the best statistician under age 40, was presented to Wong in August. It is considered to be one of the highest honors in the field.

    Wong is modest about his accomplishments. "Personally, it's very gratifying to be recognized by one's peers in one's profession, especially in the early to middle stages of one's career," he said. But, he added, laughing, "What it really means is that they want you to work harder in the future!"

    Wong, who came to the United States from Hong Kong in 1973, has been a member of the University faculty since 1980.

    One of his current projects is a collaborative effort with Chin-Tu Chen, Assistant Professor in Radiology. Chen and Wong are working on ways to improve image reconstruction from Positron Emission Tomography, or PET.

    PET is a noninvasive technique often used to map the brain to study metabolism and diagnose such illnesses as brain tumors, Alzheimer's disease and schizophrenia.

    Unlike other scanning methods, PET scans allow researchers to study physiology and biochemistry in the body. They do this, for example, by attaching a mildly radioactive isotope to glucose, a chemical that is taken up by cells that are metabolically active. The glucose is injected into the patient, and the radioactive isotope acts as a beacon, illuminating places in the body where metabolic activity is high.

    A brain tumor, for example, may metabolize glucose abnormally and show either a greater or lesser concentration of the radioisotope relative to the surrounding tissue.

    "The statistician and the medical physicist reconstruct the radioisotope density within the area that is being scanned," Wong explained. This reconstruction gives rise to the "picture" that a radiologist interprets.

    Chen said PET data have a lower resolution than images obtained from Magnetic Resonance Imaging (MRI) or Computerized Tomography (CT) scans. But Chen and Wong incorporate data from MRIs and CTs into their statistical model, improving the precision of the PET data. "That's a unique approach that we've pioneered here at U of C," Chen said.

    Despite Wong's substantial involvement in several application areas, most statisticians view his theoretical work as even more impressive. His papers on two areas of statistics in particular have become modern landmark works.

    In the area known as semiparametric inference, Wong and his students have studied how classical statistical theory (developed in the 1950s and '60s by the top statisticians of that time, including Chicago's R.R. Bahadur) could be extended from finite into infinite dimensions.

    And Wong's work on "partial likelihood" has shown how most of the statistical information in data can often be captured by modeling only a portion of the process, making the techniques of analysis much simpler and more broadly applicable.

    Stephen Stigler, Professor and former Chairman of Statistics, said, "It's remarkable how he has taken areas that others considered either 'played out' or simply too hard and has made striking advances. He has shone a light in corners that most mathematical statisticians had given up on."

    Wong himself is much more modest. He insists that his real love is just thinking about how to solve problems. "I'm really more of a theoretician than an applied scientist," he said.

    -- Diana Steele