Background
Methods
Computational modeling
The coupled neural mass model
Parameter | Description | Value (default) |
---|---|---|
t | Sample time | 0.002 s |
P(t) | The pulse density of an input signal to the excitatory population | 550 spikes/s−1 |
Noise | Random fluctuations around average level of P(t) | 1.0 |
A he(t) | Amplitude of the EPSP | 1.6 mV |
A hi(t) | Amplitude of the IPSP | 32 mV |
a he(t) | Shape parameter of EPSP | 55 s−1 |
b he(t) | Shape parameter of EPSP | 605 s−1 |
a hi(t) | Shape parameter of IPSP | 27.5 s−1 |
b hi(t) | Shape parameter of IPSP | 55 s−1 |
g | Parameter of sigmoid function that relates membrane potential to impulse density | 25 s−1 |
q | Parameter of sigmoid function that relates membrane potential to impulse density | 0.34 mV−1 |
Vd1 | Threshold potential used in the sigmoid function that relates membrane potential to impulse density for main population of excitatory neurons | 7 mV |
Vd2 | Threshold potential used in sigmoid function that relates membrane potential to impulse density for inhibitory neurons | 7 mV |
C1 | Connection strength between main population of excitatory and inhibitory neurons | 32 |
C2 | Connection strength between inhibitory neurons and main population of excitatory neurons | 3 |
S | Coupling strength between neural masses (gain factor) | 1.5 |
T | Time delay for the coupling between neural masses | 0.002 s |
N | Number of neural masses (/nodes) in network model | 78 |
Simulation of pathophysiology in early AD
Scenario | Parameter | Parameter description | Parameter value | AD-mediated pathology | ||
---|---|---|---|---|---|---|
Control condition | AD-like scenario | Contrast scenario | ||||
Pyramidal neuronal hyperactivity | ||||||
1A | Vd1 | Threshold potential of excitatory populations | 7 | 6 | 8 | A lower Vd1 value causes the excitatory populations to become hyperexcitable |
1B | he(t) function (a1 and b1) | Excitatory post-synaptic potential (EPSP) | a1: 55 b1: 605 | a1: 48 b1: 540 | a1: 62 b1: 670 | Increasing parameters a and b of the he(t) function will increase the postsynaptic excitatory amplitude and duration of both the excitatory and inhibitory populations |
1C | S | Global coupling factor | 1.5 | 2.0 | 1.0 | A higher global coupling factor results in stronger excitatory output (E(t)) multiplication and thus increased excitatory innervation of the excitatory population in the coupled neural masses |
Inhibitory neuronal dysfunction | ||||||
2A | Vd2 | Threshold potential of inhibitory populations | 7 | 8 | 6 | A higher Vd2 value causes the inhibitory populations to become hypoexcitable |
2B | hi(t) function (a2 and b2) | Inhibitory post-synaptic potential (IPSP) | a2: 27.5 b2: 55 | a2: 40 b2: 70 | a2: 17.5 b2: 35 | Higher values of parameters a and b of the hi(t) function will decrease the postsynaptic inhibitory amplitude and duration in the excitatory populations |
2C | C2 | Coupling from inhibitory to excitatory populations | 3 | 2 | 4 | A lower C2 value will decrease the inhibitory to excitatory coupling |
-
1A: (Intrinsic) pyramidal neuronal hyperexcitability. Different pyramidal neuronal excitability levels were obtained by adjusting the excitatory neuron firing threshold parameter Vd1(t). In the “healthy” or control condition, the threshold value for the excitatory (Vd1) neuronal populations had a value of 7 and this was altered to the value 6, meaning that the threshold was lower and we thus simulated a network with AD-like pyramidal neuronal hyperexcitability. The pyramidal neuron threshold potential was set to 8 to generate a contrast scenario with pyramidal neuronal hypoexcitability.
-
1B: Increased excitatory postsynaptic potential. To simulate the effect of increased extracellular glutamate levels, we changed the EPSP curve that was modeled as the impulse response function he(t) with parameters a1 and b1 (S2 Fig). Although increased neurotransmitter concentration is physiologically translated to higher EPSP frequencies in the postsynapse and not higher amplitude or duration, in this model, we regarded increased EPSP amplitude as synchronous EPSPs and thus a summation of multiple EPSPs. In control condition, a1 had a value of 55 and b1 of 605 s−1 that were set to 48 and 540 s−1 for AD-like increased EPSP amplitude and duration respectively (Table 2). Important to note here is that when we changed the EPSP curve, this not only affected the pyramidal (excitatory) neuronal population but also the inhibitory neuronal population, because both excitatory and inhibitory populations received excitatory input in the model. Changing the EPSP curve thus influenced both excitatory and inhibitory activity. The excitatory impulse response function he(t) parameters a1 and b1 received a value of 42 and 670 to simulate a contrast scenario with decreased EPSP amplitude/duration.
-
1C: Increased excitatory to excitatory coupling. As alternative scenario of increased excitatory signals in the circuit (due to glutamate reuptake block by AD pathology), we could modulate the global coupling factor (S) between coupled neural masses. We increased the S value from 1.5 in control to a value of 2.0 in the AD-like scenario, which led to a stronger multiplication of the excitatory output signal (that is spike density (E(t)) between the pyramidal neuronal populations of two coupled neural masses and thus more excitatory input to the excitatory populations only). In the non-AD-like contrast scenario, the S parameter was set to a value of 1.0 to simulate reduced excitatory input towards the neural masses.
-
2A: (Intrinsic) inhibitory neuronal hypoexcitability. Similar adjustments were made as in scenario 1A but now for the firing threshold of the inhibitory neuronal populations (parameter Vd2), i.e., the control scenario had a Vd2 value of 7, the AD-like scenario of inhibitory hypoexcitability received a Vd2 value of 8 (and thus a higher firing threshold), and the contrast-scenario received a Vd2 value of 6, reflecting inhibitory hyperexcitability.
-
2B: Decreased inhibitory post-synaptic potential. To simulate AD-like decreased inhibitory transmission and a reduced number of postsynaptic inhibitory receptors in the pyramidal neuronal population, we changed the amplitude and duration of the IPSP of the pyramidal population only (because the inhibitory neurons did not receive inhibitory input in the current model). Parameters a2 and b2 of the impulse response function hi(t) determined the IPSP shape and had a value of 27.5 and 55 s−1 in control condition (S2 Fig) and were adjusted to 40 and 70 s−1 to simulate AD-like reduced IPSP for a2 and b2 respectively. The parameters for the IPSP received a value of 17.5 and 35 s−1 to simulate a contrast scenario with increased IPSP (S2 Fig).
-
2C: Decreased inhibitory synaptic coupling. To simulate a loss of functional inhibitory synapses and therefore reduced inhibitory synaptic coupling strength, we adjusted the local inhibitory synaptic coupling coefficient C2 that determined the inhibitory to excitatory coupling strength. In control conditions, the C2 was 3, and to simulate AD-like loss of inhibition this parameter received a value of 2. The C2 parameter was increased to 4 in the non-AD-like contrast scenario to mimic stronger inhibition of excitatory synapses.
Human data
Subjects
MEG data acquisition and analyses
MEG source reconstruction
Outcome measures
Combining simulated and human data
Statistical analyses
Results
Pyramidal neuronal hyperactivity
Scenario 1A: (Intrinsic) pyramidal neuronal hyperexcitability
Scenario 1B: Increased excitatory postsynaptic potential
Scenario 1C: Increased excitatory to excitatory coupling
Inhibitory neuronal dysfunction
Scenario 2A: Decreased inhibitory interneuron excitability
Scenario 2B: Decreased inhibitory postsynaptic potential
Scenario 2C: Decreased inhibitory synaptic coupling
Simulated MEG power spectra
Human MEG
SCD− | MCI+ | |
---|---|---|
N | 18 | 18 |
Female/male | 10/8 | 9/9 |
Age (y) | 64.2 (± 6.1) | 64.1 (± 6.2) |
MMSE score | 27.8 (± 2.1) | 25.8 (± 1.9)** |
Combining simulated and human MEG data
Human AD | AD-like scenarios | Contrast scenarios | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1A | 1B | 1C | 2A | 2B | 2C | 1A | 1B | 1C | 2A | 2B | 2C | ||
Parameter | Vd1 | EPSP | S | Vd2 | IPSP | C2 | Vd1 | EPSP | S | Vd2 | IPSP | C2 | |
Direction | up | up | up | down | down | down | down | down | down | up | up | up | |
Neuronal activity | n.a. | Higher | Lower | Higher | Higher | Higher | Higher | Lower | Higher | Lower | Lower | Lower | Lower |
Oscillatory behavior | Slower | Slower | Slower | Slower | Slower | Faster | Slower | Faster | Faster | Faster | Faster | Slower | Faster |
Total power | Higher (n.s.) | Higher | Higher | Higher | Higher | Higher | Higher | Lower | Lower | Lower | Lower | Lower | Lower |