Why Functional magnetic resonance imaging Is Not the Best Way to Study Brain Function
I believe that functional magnetic resonance imaging (fMRI) is based on questionable science.
Let’s start by analyzing the basic concepts behind fMRI and discussing why these concepts are questioned by many in the scientific community.
Concepts behind fMRI
Concept One: There are well-defined networks of brain cells: the dopamine network, the serotonin network, and others. Each network is composed of multiple members connected together by axons (tiny electric wires) and works sequentially and independently from other networks. A signal starts in a cell at one end and spreads to the next member cell, and the next, and so on. This is like a football game where player A catches the ball, uses energy to catch the ball and to throw it to Player B, who also increases his metabolism to catch the ball and send it to the next player, and so on. Each cell will increase its activity to respond to the signal from the previous cell member and to send it to the next cell member of that well-defined network. This is known as cell activation.
Concept Two: These networks are inactive at rest. This is called default mode, and the network is called the default mode network, DMN for short. This is a resting, fixed, low-level metabolic activity against which we measure all other activities. This is like comparing your metabolism while asleep to when awake. We should expect the metabolic activity to lessen during sleep and increase with waking-state activities. Thus, with mental activity, this default-mode metabolism increases in actively working cells—the cells are metabolically active or “activated.”
Concept three: Brain cells don’t store any oxygen. As the brain cell becomes more active, it needs more energy, thus more oxygen. The blood flow increases in the cell’s active area in response to the increased metabolism to provide needed oxygen. The oxygen is taken from oxyhemoglobin (oxygen attached to hemoglobin) in the blood. This converts oxyhemoglobin to deoxyhemoglobin (de = take away, oxy = oxygen). With the loss of oxygen (two negative charges), deoxyhemoglobin becomes positively charged. In a magnetic field, positive and negative particles move in opposite directions in relation to the magnetic pole. This change of direction is detected in the susceptibility images. We can generate pictures of any magnetically charged chemicals in the brain, based on their movements in relationship to the magnet. The susceptibility image could be color-coded with any colors desired. This image could be superimposed on the routine MRI image, in any color your heart desires, creating the fMRI images. This generated signal is called the BOLD (blood oxygen level-dependent) signal. This signal reflects the decrease in oxygen from its use by active cells. Last, please note that fMRI doesn’t prove that a certain chemical is released, but it assumes so, based on the type of cells present at a specific location in the brain.
As we now understand the basic concepts behind fMRI, we can move on to analyze them one by one.
Problems with fMRI concepts
Concept One Problems: Networks are independent of each other. Neurologists know that all systems are dynamically balanced and connected all of the time. If a patient has Parkinson’s disease due to a reduced release of dopamine, they can treat him by increasing his dopamine brain release. However, if the patient can’t tolerate the dopamine-increasing medications, they can get the same effect by blocking a different network: the acetylcholine network. We get the same result by affecting either system. Genetic studies of schizophrenia (supposedly a dopamine disorder) found disordered genes related to glutamate metabolism and not dopamine. Therefore, a problem in the same system can be caused by a problem in one chemical that causes imbalance of another. Here, glutamate causes an imbalance of dopamine. The systems do not work independent of each other. Each network is composed of multiple members connected together by axons (tiny electric wires) and works sequentially. For example, the connections between these centers are not yet developed in children at eight years of age because the axons are as yet myelinated, like copper wires without the plastic coating will cause short wiring and no electricity will travel down the wire. However, these children are able to do the same type of thinking and feeling as adults by using the centers without the connections between them. Those are the connections that they presumable generates the fMRI signal. Their fMRI images look similar to fMRI images in adults whose connections between centers are developed. This questions the concept of the default mode network, the activity of which, its proponents claim, is based on connections between the cells in the network.
Concept Two Problems: The baseline activity is that which reflects the network’s basic metabolic energy needs. With activity, these networks increase their metabolic needs and consume more oxygen, which fMRI believers call activation. They call the decrease in oxygen consumption below the normal resting level “deactivation”. We have evidence that the default networks consume more energy at rest than while doing a range of explicit tasks. This should decrease the oxygen consumption, causing less blood-oxygen-level dependent signal to reveal itself on an fMRI, not an increase (activation). If the resting state of the default mode network is the baseline of metabolic function, how can the brain function with less energy? Studies on energy consumption of the brain show minimal changes to task-related energy consumption, too small to be accurately measured.
Concept Three Problems: The BOLD (blood oxygen level-dependent) signal reflects the decrease in oxygen levels due to its use by active cells. It’s proven that merely changing one’s breathing pattern, without doing any mental activity at all, changes the fMRI blood oxygen level-dependent signals, both baseline metabolic activity and the metabolic activity associated with mental tasks, or activation tasks. Do you feel more love for your mate when you breathe faster or when you breathe slower? Or are you like most of us, never feeling any difference in your love by changing your breathing pattern? It appears that the fMRI signal is dependent on blood flow and not on metabolic activity in the cell. We change the brain blood flow in comatose patients by changing the breathing pattern on their ventilators. We know that certain locations in the brain are more susceptible to certain pathology than others; for example, a drop in oxygen causes more brain injury in certain locations than others. Could the difference in signal be from the difference in normal cell metabolic activity? Thus, could the default mode be associated with the normal variation of metabolic activity?
fMRI images measure an average of twenty seconds of activity. Does our brain function that slowly? Do you spend twenty seconds to just look at an image? Most people review slideshows at the rate of three to four seconds for each image. Are we including another sixteen seconds of other activities in the brain? Which of the twenty seconds are we actually measuring?
The Problem with “The Brain at Rest Consumes Less Energy” Concept: You should expect baseline metabolic needs to stay the same all the time. If you put people under anesthesia and stimulate their vision with lights, you get a larger blood oxygen level-dependent signal, while there is no increase in brain-cell activity because they’re unconscious. Under anesthesia, you should expect a decrease in the BOLD signal. Many question the concept “The brain does nothing unless it’s instructed to.” That’s like saying that we do no work unless we receive a bill, and only then will we start working. If this was the case, nobody would ever have savings. We know without doubt that all neurotransmitters are premade, and stored in nerve terminals (in vesicles). These neurotransmitters are released when the electric signal reaches the nerve terminal. The neurotransmitters aren’t made the moment we need them. Just like earning money, you work all the time, but you save the money and spend it only when you need to; we consume more energy working than spending money. This matches the brain’s increased energy consumption at rest and less consumption with activity.
It’s been proven that many decreases in function, which fMRI enthusiasts call deactivation, happen independent from the task. Four different areas have shown this deactivation, regardless of whether or not a mental task is being performed. Areas of activation and inactivation in the brain aren’t consistent in all studies for the same task. fMRIs that measure oxygen use, and PET scans that measure glucose metabolism, show identical images to task activation. However, there’s no evidence that the oxygen consumption has any temporal relationship to glucose consumption. The signal in fMRI must have another source, probably a neurophysiologic and not a neurocognitive (thinking) source. If an fMRI signal suggests an internal thought process, why do we do the same when we pay passive attention to the external environment? We still get activation and deactivation signals. In a study on rats’ brain signals, rats showed the changes of activation and deactivation as they navigated a maze. However, when they stopped the navigation and mental work, the activation and deactivation continued even though they weren’t doing any mental processing!
Further criticism: The fMRI images use subtraction of the activation signal from baseline activity. Flaws and fallacies in subtraction have been debated by many scientists. The international radiology encyclopedia Radiopaedia.org states, “fMRI is technically challenging to perform as the techniques used to visualize cortical activity (typically BOLD imaging) rely on minute changes in a low signal-to-noise ratio (SNR) environment.” Basically they’re saying that it’s difficult and unreliable to try to listen to a whisper in a very noisy environment. You can use this type of imaging to check the surface of the brain, the cortex, but not the deep structures in the brain, such as the ventral tegmental area. The continuous recording of OEF (oxygen extraction fraction, the amount of oxygen taken from blood) continuously fluctuates, suggesting that these cells are in a dynamic state and not a fixed baseline.
Be careful when you read about brain chemistry of love if it is based on studies using fMRI.
What you hear or read might not be based on real science.
Fred Nour, MD