Science & Fortuity

The historical account below is published with the author's permission. I annotated two typographical errors. The document is best read with fullscreen view. The reader may find this option at the right-hand corner of the task bar at the bottom of the text window or may open a new window, using this link. 

The lecture revisits the development of the fluorodeoxyglucose method for the imaging of the brain's metabolic activity. The author, whose contribution was instrumental in the endeavor, presented the lecture via video at Brookhaven National Laboratory (BNL) Oct. 19, 2012. The occasion was a symposium convened to celebrate the designation of BNL's chemistry department by the American Chemical Society as a historical site in recognition of its pivotal role in the adaptation of the method to the non-invasive use in humans.

The history of the fluorodeoxyglucose method represents a striking example for the value of basic research. It demonstrates that the translation of discoveries from bench to bedside involves unrelenting dedication and, at times, serendipity. Furthermore, this example shows that the process may require decades of continued funding, before the investment eventually comes to fruition. None of the investigators who embarked on this endeavor in the 1940s fathomed that their efforts would lead one day to the most widely used diagnostic tool for the staging of cancer anywhere in the body.

Lou Sokoloff BNL FDG Symp Lecture 2012

Absolutes, Relatives, Brain Imaging & Steroids

The functional brain imaging methods most commonly used in humans today are functional magnetic resonance imaging (fMRI) using blood oxygen-level dependent (BOLD) signals and the subtractive water method with positron emission tomography (PET). Both procedures record changes in local cerebral blood flow from a baseline. Local cerebral blood flow is associated with energy demands of activated nerve cells. The cells consume glucose sugar and oxygen to process information which cannot be stored in the brain and must be supplied on demand with the blood stream. Hence, blood flow increases with increased nerve cell activity. In the simplest conception of both procedures known as block design, measurements acquired over several minutes of mental activation are compared with measurements acquired during equivalent epochs of rest. The statistically significant difference is considered the result of cerebral activation.

In conventional fMRI, blood oxygen-level dependent signals are used to detected changes in blood flow (Ogawa and others, 1993). With the subtractive water method (Fox and others, 1984), water labeled with ¹⁵O, a radioactive positron-emitting isotope of oxygen, is used as tracer that freely diffuses into the brain tissue. The local concentration of the tracer is commensurate with local blood flow and can be imaged in a PET scanner. However, calibration for absolute blood flow is wrought with difficulty and has not found wide-spread application. Because the blood flow is not quantified, the differences between the compared mental states remain relative. That is, cerebral activation is usually expressed in percent difference from the state used as reference or in units of statistically significant difference (statistical parametric mapping).

By contrast, the cerebral glucose consumption is more directly related to nerve cell activity than cerebral blood flow. The deoxyglucose method of Sokoloff and others (1977) permits us to measure the local cerebral rate of glucose utilization. Deoxyglucose is an analogue to glucose that accumulates in the brain tissue commensurate with glucose consumption. Tagging deoxyglucose with ¹⁸F, a radioactive positron-emitting isotope of fluorine, the tracer's accumulation in the brain can be imaged with PET. The [¹⁸F]fluorodeoxyglucose method, therefore, provides a snapshot of the brain's energy consumption (Reivich and others, 1979). Although this snapshot needs 90 minutes to develop because of the tracer kinetics involved, the procedure constitutes an indispensable tool for the detection of long-term, pseudo-stationary changes in absolute cerebral metabolic activity as a consequence of disease or trauma. Below, I discuss one example.

After the collapse of the regime of Nicolae Ceauşescu at the end of 1989, U.S. parents began to adopt children from Romanian orphanages. The children had been kept in circumstances of great depravity, producing profound behavioral problems similar to autism (American RadioWorks report, 2006). Visiting scientists reported behavioral patterns resembling those the eminent American psychologist Harry Harlow had so aptly described in primates raised in isolation and with surrogates.

Although the adoptees were brought to the U.S. at very young age, some developed cognitive and behavioral differences, including impulsive reactions as well as attention and social deficits, in the years after their arrival.

Research at the orphanages provided evidence that the children had persistently augmented levels of cortisol n their blood stream as a result of the severe stress they endured (Carlson and Earls, 1997). Cortisol is a known steroid stress hormone produced in the adrenal glands and can fundamentally affect brain maturation. The hormone suppresses the activity of glia. A type of glia, astrocytes, helps regulate the extracellular glutamate concentration. Glutamate constitutes the most prevalent excitatory neurotransmitter in the brain, playing a major role in the stabilization of connections between nerve cells during brain maturation. Elevated concentrations of extracellular glutamate can trigger pre-programmed cell death known as apoptosis, otherwise occurring only during early stages of brain development. Presumably, the orphans' excessive stress-related exposure to cortisol led to modifications of nerve cell networks, underlying the children's behavioral differences. Imaging the brain's energy consumption provided a method to uncover whether and where nerve cell activity changed in cerebral cortex as a consequence of the children's stay in the orphanages. 

Using the fluorodeoxyglucose method, Chugani and others (2001) could show that the use of glucose was drastically reduced in the cerebral cortex of the orphans enrolled in the study, particularly in temporal and prefrontal cortical areas and in structures of the limbic system, notably the amygdala. The cortical regions are involved in executive functions and short-term memory crucial for social behavior and affect. The amygdala play an important role in fearful reactions. The observed reductions in energy consumption could not have been detected with the standard fMRI or PET procedures discussed above. The fluorodeoxyglucose method, hence, constitutes the procedure of choice when the fundamental metabolic state of the brain is in question.

Addendum

• Take some time and listen to this show on National Public Radio's This American Life with the title "Unconditional Love". The first half of the show is about an orphaned Romanian boy adopted by an American couple at the age of eight. It demonstrates in great clarity the at times overwhelming difficulties the family faced to remedy important steps of personality development that were missed early in the boy's life. Finally, the challenges were overcome with passion and a professional attitude. It is reassuring to find out that success is possible (10/23/10).

References

• Carlson M, Earls F (1997) Psychological and neuroendocrinological sequelae of early social deprivation in institutionalized children in Romania. Ann N Y Acad Sci 807:419-428.
• Chugani HT, Behen ME, Muzik O, Juhász C, Nagy F, Chugani DC (2001) Local brain functional activity following early deprivation: a study of postinstitutionalized Romanian orphans. Neuroimage 14:1290-1301.
• Fox PT, Mintun MA, Raichle ME, Herscovitch P (1984) A noninvasive approach to quantitative functional brain mapping with H2 (15)O and positron emission tomography. J Cereb Blood Flow Metab 4:329-333.
• Ogawa S, Menon RS, Tank DW, Kim SG, Merkle H, Ellermann JM, Ugurbil K (1993) Functional brain mapping by blood oxygenation level-dependent contrast magnetic resonance imaging. A comparison of signal characteristics with a biophysical model. Biophys J 64:803-812.
• Reivich M, Kuhl D, Wolf A, Greenberg J, Phelps M, Ido T, Casella V, Fowler J, Hoffman E, Alavi A, Som P, Sokoloff L (1979) The [18F]fluorodeoxyglucose method for the measurement of local cerebral glucose utilization in man. Circ Res 44:127-137.
• Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, Pettigrew KD, Sakurada O, Shinohara M (1977) The [14C]deoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J Neurochem 28:897-916.

Gustav Fechner & Functional Brain Imaging

Today we celebrate the birthday of Gustav Theodor Fechner. He was born in 1801. Anecdotes have it that he was a late-night owl and that, as a young student, he liked to stay in bed daydreaming in the mornings and was notoriously tardy even for ten-o'clock c.t. lectures (Boring EG, 1970). In the German academic system, the suffix c.t. stands for cum tempore and means that the lecture actually begins fifteen minute past the hour.

He must have spent his time with the right thoughts. At the age of 34, he had completed his M.D. and Ph.D. degrees and was tenured as full professor on his way to become one of the first and most eminent experimental psychologists. He believed that the physical world and our perception of it were complementary and set out to find a mathematical relationship between the physical world and its mental image.

While Fechner began to explore methods to quantify mental work, he became aware of Ernst Heinrich Weber's studies. Weber had asked blindfolded participants to compare packages of differing weight placed in each hand. He observed that the smallest weight that could be judged as different changed in fixed proportion with the weight of reference. The heavier the reference weight, the greater the difference in weight had to be to be noticed. The reported change in sensation Δs was a function of the ratio between the difference in weight Δw and the reference weight W multiplied by a factor r:

(1)    f(Δs) = r x  Δw/W

Integrating function (1) and substituting weight for stimulus R results in:

(2)    f(S) = r x ln(R) + S_min

where sensation S is a function of the natural logarithm of stimulus R multiplied by factor r plus a constant S_min which constitutes the smallest noticeable difference in R. Hence, sensation S increases in direct proportion with the logarithm of stimulus R with the slope of factor r. Consequently, stimulus magnitude had to be doubled for the increase to be noticed.

In reverence of Weber's work, Fechner called the direct proportionality between sensation and the logarithm of stimulus strength Weber's law, proposing that the relationship may apply to all senses. Confirming his hypothesis, he could provide evidence that Weber's law adequately described the relationship between perceived luminosity and the brightness of stars. He believed that the law might open ways to indirectly measure mental processes, trailblazing a science of quantifying mental activity. Alas mind work remained inferred from the participants' judgment. In acknowledgment of the pioneering contributions to experimental psychology of both men, the law is known as Weber-Fechner law today.

One hundred-fifty years later, the direct measurement of mental processes remains elusive. However, methods have been found to measure cerebral work. Brain cells utilize glucose and oxygen to process information. Both have to be supplied by blood flow on demand, since they cannot be stored in the brain tissue. Thus, local cerebral blood flow increases with the increase in nerve cell activity.

Since the 1940's, Seymour Kety, Lou Sokoloff, and others were working out methods to measure local cerebral blood flow with radioactive tracers, initially at the University of Pennsylvania and later at the National Institutes of Health (Sokoloff, 2000). Around 1970, Lou Sokoloff and others succeeded in developing a method with which the rates of local cerebral glucose utilization could be determined in animals (Sokoloff and others, 1977).

With the advent of Positron Emission Tomography (PET), both methods became widely used in humans. Since then, numerous studies have demonstrated that brain work and blood flow indeed increase in logarithmic proportion with the magnitude of stimulation. Electrical nerve cell discharges known as spikes encode stimulus magnitude. My colleagues and I observed with micro-electrode recordings from nerve cells in the cerebral cortex that the cells are limited in their ability to increase the spike rate with increasing frequency of stimulation. Instead, they dynamically scale down their responsiveness with increasing frequency, while continuing to respond time-locked to the stimuli (Melzer and others, 2006). This ability, known as gain control, may underlie the logarithmic relationship between sensory stimulus, brain cell response and mental perception.

Combining PET with functional magnetic resonance imaging (fMRI), Dettmers and others (1996) were able to show a tight association between stimulus-related increases in local cerebral blood flow measured with PET and the blood oxygen level-dependent (BOLD) signal measured with fMRI. Hence, modern non-invasive tomographic brain imaging methods support that the law of Weber and Fechner may apply to brain work. This is not exactly the outcome Professor Fechner had in mind, but he would certainly find the findings exciting, were he still alive today.

References

• Boring EG (1970) History of Experimental Psychology. Appleton, New York.
• Dettmers C, Connelly A, Stephan KM, Turner R, Friston KJ, Frackowiak RS, Gadian DG (1996) Quantitative comparison of functional magnetic resonance imaging with positron emission tomography using a force-related paradigm. Neuroimage 4:201-209.
• Melzer P, Sachdev RN, Jenkinson N, Ebner FF (2006) Stimulus frequency processing in awake rat barrel cortex. J Neurosci 26:12198-12205.
• Sokoloff L (2000) Seymour S. Kety. Biographical Memoirs. The National Academies Press. Wash. D.C.
• Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, Pettigrew KD, Sakurada O, Shinohara M (1977) The [14C]deoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J Neurochem 28:897-916.

Imaging Discord in the Brain

The advent of functional brain imaging has revolutionized the fashion in which psychologists and psychiatrists look at the brain. The pictures of behavior-related cerebral activation provide unprecedented information leading to new hypotheses about the workings of our mind.

However, it is of utmost importance to keep in mind that non-invasive functional brain imaging methods do not allow us to record nerve cell activity directly. With positron emission tomography (PET), single photon emission computed tomography (SPECT), and functional magnetic resonance imaging (fMRI), the most-frequently used procedures detect changes in local cerebral blood flow while the participants are exposed to sensory stimulation or execute tasks. Brain cells need sugar and oxygen to fuel the chemical reactions necessary for information processing. Both resources cannot be stored in the brain and thus have to be delivered on demand. Hence, local blood flow increases when nerve cells are activated, resulting in a tight association between nerve cell activity and blood flow under normal physiological conditions.

The molecular mechanisms that couple blood flow to nerve cells activity are not yet fully understood. Glutamate constitutes the predominant excitatory neurotransmitter in the cerebral cortex. This neurotransmitter, its precursors and metabolites as well as its cellular receptors may play a crucial role in the coupling of the two events. However, molecules unrelated to glutamate may also be important. Nitric oxide (NO) and adenosine are known to influence the blood flow response.

In addition to our lack of knowledge on the coupling between the nervous and the vascular response, blood flow measurements inherently cover a volume of brain tissue and do not permit us to identify precisely which nerve cells drive the observed change in flow.

In this week's issue of the journal Nature, Kerri Smith informs us on new findings relevant to the interpretation of functional brain imaging. Yevgeniy Sirotin and Aniruddha Das demonstrate the consequences of the uncertainties discussed above in a letter to Nature entitled "Anticipatory haemodynamic signals in sensory cortex not predicted by local neuronal activity" (Nature 457:475-479). The authors used optical imaging for the fine-grain mapping of changes in blood flow in exposed primary visual cortex of monkeys. The animals were trained to react to a small visual cue. As anticipated, blood flow increased locally in visual cortex after stimulus onset, and the researchers could record concomitantly increased nerve cell activity with wire electrodes inserted into the brain tissue at this location.

Remarkably, blood flow also increased, when the monkeys expected the visual cue to appear, but it was not presented. The anticipation alone was sufficient to significantly increase the local blood flow. By contrast, Sirotin and Das were not able to detect any increase in nerve cell activity that could be related to the anticipatory increase in blood flow.

The apparently discordant findings may not be entirely surprising. The monkeys were accustomed to treats as reward for their participation. Their readiness for the task may have activated neuromodulatory inputs to visual cortex that remain sub-threshold under ordinary conditions and do not trigger nerve cell activity directly, but facilitate the nerve cell response to the imminent stimulus. How such sub-threshold nerve cell signals may increase local blood flow remains an open question.

The discrepancy between blood flow and nerve cell activity Sirotin and Das observed suggests that blood-flow based brain imaging data must be considered with utter prudence, when complex behaviors are examined that involve the subjects' active participation and anticipation. The findings should caution those who strive to correlate patterns of cerebral blood flow with socio-affective mental disorders and criminality in the hope of developing novel predictors for our actions.

Neurolaw is an attempt to associate patterns of brain activity with criminal behavior. Terry Gross interviewed the eminent American neuroscientist Michael Gazzaniga on this issue on National Public Radio's Fresh Air broadcast July 28, 2008. I once wrote down my thoughts on this idea in secret ink. If you wish to spare a few minutes, click on the video, let the magic unfold and enjoy!